(Vasile Cîrtoaje) Mathematical Inequalities Vol 2

Vasile Cîrtoaje
C j
L
MATHEMMATICAL
.M
INEQUA
Q ALITIES
D
A
Volume 2
PI
M
LY
SYMM
METRIC
O
RATIONAL AND NONRATIONAL

RATIONAL AND
.M
ALITIES
INEQUA
W
W
W
UNIVERSITY OF PLLOIESTI, ROMANIA

2015
W
W
W
.M
O
LY
M
PI
A
D
.M
L
Contents
L
1 Symmetric Rational Inequalities 1
.M
1.1 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
D
2 Symmetric Nonrational Inequalities 257
A
2.1 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
2.2 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
PI
3 Symmetric Power-Exponential Inequalities 419
M
3.1 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
LY
3.2 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
4 Bibliography 473
O
.M
W
W
W
i
ii Vasile Cîrtoaje
L
D
.M
A
PI
M
LY
O
.M
W
W
W
Chapter 1
Symmetric Rational Inequalities
L
.M
1.1 Applications
D
A
1.1. If a, b are nonnegative real numbers, then
1 1
PI
1
+ ≥ .
M
(1 + a)2 (1 + b)2 1 + ab
LY
1.2. If a, b, c are nonnegative real numbers, then

O
a2 − bc b2 − ca c2 − a b
.M
+ + ≥ 0.
3a + b + c 3b + c + a 3c + a + b
W
1.3. If a, b, c are positive real numbers, then

W
4a2 − b2 − c 2 4b2 − c 2 − a2 4c 2 − a2 − b2
+ + ≤ 3.
W
a(b + c) b(c + a) c(a + b)
1.4. Let a, b, c be nonnegative real numbers, no two of which are zero. Prove that
1 1 1 3
(a) + 2 + 2 ≥ ;
a2 + bc b + ca c + a b a b + bc + ca
1 1 1 2
(b) + 2 + 2 ≥ .
2a2 + bc 2b + ca 2c + a b a b + bc + ca
1 1 1 2
(c) + 2 + 2 > .
a2 + 2bc b + 2ca c + 2a b a b + bc + ca
1
2 Vasile Cîrtoaje
a(b + c) b(c + a) c(a + b)

+ 2 + 2 ≥ 2.
a2 + bc b + ca c + ab
a2 b2 c2 a b c
+ + ≥ + + .
b +c
2 2 c +a
2 2 a +b
2 2 b+c c+a a+b
L
1.7. Let a, b, c be positive real numbers. Prove that
.M
1 1 1 a b c
+ + ≥ 2 + 2 + 2 .
D
b+c c+a a+b a + bc b + ca c + a b
A
PI
M
1 1 1 2a 2b 2c
+ + ≥ 2 + 2 + 2 .
b+c c+a a+b 3a + bc 3b + ca 3c + a b
LY
O

.M
a b c 13 2(a b + bc + ca)
(a) + + ≥ − ;
b+c c+a a+b 6 3(a2 + b2 + c 2 )
W
3 p a b + bc + ca

a b c
(b) + + − ≥ ( 3 − 1) 1 − 2 .
b+c c+a a+b 2 a + b2 + c 2
W
W

2
1 1 1 a+b+c

+ + ≤ .
a2 + 2bc b2 + 2ca c 2 + 2a b a b + bc + ca
a2 (b + c) b2 (c + a) c 2 (a + b)
+ 2 + 2 ≥ a + b + c.
b2 + c 2 c + a2 a + b2
Symmetric Rational Inequalities 3
a2 + b2 b2 + c 2 c 2 + a2 3(a2 + b2 + c 2 )
+ + ≤ .
a+b b+c c+a a + b + c)

1 1 1 9
+ 2 + 2 ≥ .
a2 + ab + b 2 b + bc + c 2 c + ca + a 2 (a + b + c)2
L
.M
a2 b2 c2 1
+ + ≤ .
D
(2a + b)(2a + c) (2b + c)(2b + a) (2c + a)(2c + b) 3
A
PI
M
P a 1
(a) ≤ ;
(2a + b)(2a + c) a+b+c
LY
a3 1
O
P
(b) ≤ .
(2a + b )(2a + c )
2 2 2 2 a+b+c
.M

W
X 1 1 2
≥ + .
(a + 2b)(a + 2c) (a + b + c)2 3(a b + bc + ca)
W
W
1 1 1 4
(a) + + ≥ ;
(a − b)2 (b − c)2 (c − a)2 a b + bc + ca
1 1 1 3
(b) + 2 + 2 ≥ ;
a2 − ab + b 2 b − bc + c 2 c − ca + a 2 a b + bc + ca
1 1 1 5
(c) + 2 + 2 ≥ .
a2 +b 2 b +c 2 c +a 2 2(a b + bc + ca)
4 Vasile Cîrtoaje
1.18. Let a, b, c be positive real numbers, no two of which are zero. Prove that
(a2 + b2 )(a2 + c 2 ) (b2 + c 2 )(b2 + a2 ) (c 2 + a2 )(c 2 + b2 )

+ + ≥ a2 + b2 + c 2 .
(a + b)(a + c) (b + c)(b + a) (c + a)(c + b)
1.19. Let a, b, c be positive real numbers such that a + b + c = 3. Prove that

1 1 1
+ 2 + 2 ≤ 1.
a2 +b+c b +c+a c +a+b
L
1.20. Let a, b, c be nonnegative real numbers such that a + b + c = 3. Prove that
.M
a2 − bc b2 − ca c 2 − a b
+ 2 + 2 ≥ 0.
a2 + 3 b +3 c +3
D
A
PI
1 − bc 1 − ca 1 − a b
M
+ + ≥ 0.
5 + 2a 5 + 2b 5 + 2c
LY

O
1 1 1 3
.M
+ 2 + 2 ≤ .
a2 + b +2 b +c +2 c +a +2 4
2 2 2
W

W
1 1 1 1
+ + ≤ .
4a2 + b2 + c 2 4b2 + c 2 + a2 4c 2 + a2 + b2 2
W

bc ca ab
+ 2 + 2 ≤ 1.
a2 +1 b +1 c +1

bc ca ab 1
+ + ≤ .
a+1 b+1 c+1 4

1 1 1 3
+ + ≤ .
a(2a + 1) b(2b + 1) c(2c + 1) 11a bc
2 2 2

1 1 1
+ + ≤ 1.
a3 + b + c b3 + c + a c 3 + a + b
L
.M
a2 b2 c2
+ + ≥ 1.
1 + b3 + c 3 1 + c 3 + a3 1 + a3 + b3
D
A
PI
1 1 1 3
M
+ + ≤ .
6 − a b 6 − bc 6 − ca 5
LY

O
1 1 1 1
+ 2 + 2
.M
≤ .
2a2 + 7 2b + 7 2c + 7 3
W
1.31. Let a, b, c be nonnegative real numbers such that a ≥ b ≥ 1 ≥ c and a + b + c = 3.

Prove that
W
1 1 1 3
+ 2 + 2 ≤ .
a +3 b +3 c +3 4
2
W

1 1 1 3
+ + ≥ .
2a2 + 3 2b2 + 3 2c 2 + 3 5
1.33. Let a, b, c be nonnegative real numbers such that a ≥ 1 ≥ b ≥ c and a + b + c = 3.

Prove that
1 1 1
+ 2 + 2 ≥ 1.
a +2 b +2 c +2
2
6 Vasile Cîrtoaje
1.34. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. Prove that
1 1 1 a+b+c 3
+ + ≥ + .
a+b b+c c+a 6 a+b+c
1 1 1 3
+ 2 + 2 ≥ .
a2 +1 b +1 c +1 2
L
.M
1.36. Let a, b, c be positive real numbers such that a b + bc + ca = 3. Prove that
a2 b2 c2
D
+ + ≥ 1.
a2 + b + c b2 + c + a c 2 + a + b
A
PI
M
bc + 4 ca + 4 a b + 4 bc + 2 ca + 2 a b + 2
LY
+ 2 + 2 ≤3≤ 2 + + .
a +4 b +4 c +4
2 a + 2 b2 + 2 c 2 + 2
O
p
.M
1.38. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. If k ≥ 2 + 3,

then
1 1 1 3
+ + ≤ .
W
a+k b+k c+k 1+k

W
1.39. Let a, b, c be nonnegative real numbers such that a2 + b2 + c 2 = 3. Prove that

W
a(b + c) b(c + a) c(a + b)

+ + ≤ 3.
1 + bc 1 + ca 1 + ab
1.40. Let a, b, c be positive real numbers such that a2 + b2 + c 2 = 3. Prove that
a2 + b2 b2 + c 2 c 2 + a2
+ + ≤ 3.
a+b b+c c+a
ab bc ca 7
+ + + 2 ≤ (a + b + c).
a+b b+c c+a 6
1 1 1 3
(a) + + ≤ ;
3 − a b 3 − bc 3 − ca 2
1 1 1 3
L
(b) p +p +p ≤p .
6 − ab 6 − bc 6 − ca 6−1
D
.M
A
1 1 1 3
+ + ≥ .
1 + a5 1 + b5 1 + c 5 PI
2
M
1.44. Let a, b, c be positive real numbers such that a bc = 1. Prove that
LY
1 1 1
+ + ≥ 1.
O
a2 + a + 1 b2 + b + 1 c 2 + c + 1
.M

W
1 1 1
+ 2 + 2 ≤ 3.
a2 −a+1 b − b+1 c −c+1
W
W
3+a 3+ b 3+c
+ + ≥ 3.
(1 + a)2 (1 + b)2 (1 + c)2
7 − 6a 7 − 6b 7 − 6c
+ + ≥ 1.
2 + a2 2 + b2 2 + c 2
8 Vasile Cîrtoaje
a6 b6 c6
+ + ≥ 1.
1 + 2a5 1 + 2b5 1 + 2c 5

a b c 1
+ + ≤ .
a2 + 5 b2 + 5 c 2 + 5 2
L
.M
1 1 1 2
+ + + ≥ 1.
(1 + a)2 (1 + b)2 (1 + c)2 (1 + a)(1 + b)(1 + c)
D
A
1.51. Let a, b, c be nonnegative real numbers such that PI
1 1 1 3
M
+ + = .
a+b b+c c+a 2
LY
Prove that
3 2 1
≥ + 2 .
a+b+c a b + bc + ca a + b2 + c 2
O
.M
1.52. Let a, b, c be nonnegative real numbers such that
7(a2 + b2 + c 2 ) = 11(a b + bc + ca).

W
Prove that
W
51 a b c
≤ + + ≤ 2.
28 b+c c+a a+b
W
1 1 1 10
+ 2 + 2 ≥ .
a2 +b 2 b +c 2 c +a 2 (a + b + c)2
1 1 1 3
+ 2 + 2 ≥ .
a2 − ab + b 2 b − bc + c 2 c − ca + a 2 max{a b, bc, ca}
a(2a + b + c) b(2b + c + a) c(2c + a + b)

+ + ≥ 6.
b2 + c 2 c 2 + a2 a2 + b2
a2 (b + c)2 b2 (c + a)2 c 2 (a + b)2

+ 2 + ≥ 2(a b + bc + ca).
b2 + c 2 c + a2 a2 + b2
L
1.57. If a, b, c are real numbers such that a bc > 0, then
.M
1 1 1

X a a b c
3 +5 + + ≥8 + + .
b2 − bc + c 2
D
bc ca a b a b c
A
PI
M
1 1 1

(a) 2a bc + + + a2 + b2 + c 2 ≥ 2(a b + bc + ca);
a+b b+c c+a
LY
a2 b2 c2 3(a2 + b2 + c 2 )
(b) + + ≤ .
O
a+b b+c c+a 2(a + b + c)

.M
W
a2 − bc b2 − ca c 2 − a b 3(a b + bc + ca)
(a) + 2 + + ≥ 3;
b2 + c 2 c + a2 a2 + b2 a2 + b2 + c 2
W
a2 b2 c2 a b + bc + ca 5
W
(b) + + + 2 ≥ ;
b2 + c 2 c 2 + a2 a2 + b2 a + b2 + c 2 2
a2 + bc b2 + ca c 2 + a b a b + bc + ca
(c) + 2 + 2 ≥ 2 + 2.
b +c
2 2 c +a 2 a +b 2 a + b2 + c 2
a2 b2 c2 (a + b + c)2
+ + ≥ .
b2 + c 2 c 2 + a2 a2 + b2 2(a b + bc + ca)
10 Vasile Cîrtoaje
2a b 2bc 2ca a2 + b2 + c 2 5
+ + + ≥ .
(a + b)2 (b + c)2 (c + a)2 a b + bc + ca 2
ab bc ca 1 a b + bc + ca
+ + + ≥ 2 .
(a + b)2 (b + c)2 (c + a)2 4 a + b2 + c 2
L
.M
3a b 3bc 3ca a b + bc + ca 5
D
+ + ≤ 2 + .
(a + b)2 (b + c)2 (c + a)2 a + b2 + c 2 4
A
PI
M
a3 + a bc b3 + a bc c 3 + a bc
LY
(a) + + ≥ a2 + b2 + c 2 ;
b+c c+a a+b
O
a3 + 2a bc b3 + 2a bc c 3 + 2a bc 1
(b) + + ≥ (a + b + c)2 ;
b+c c+a a+b 2
.M
a3 + 3a bc b3 + 3a bc c 3 + 3a bc
(c) + + ≥ 2(a b + bc + ca).
b+c c+a a+b
W
W
W
a3 + 3a bc b3 + 3a bc c 3 + 3a bc
+ + ≥ a + b + c.
(b + c)2 (c + a)2 (a + b)2
a3 + 3a bc b3 + 3a bc c 3 + 3a bc 3
(a) + + ≥ ;
(b + c)3 (c + a)3 (a + b) 3 2
3a3 + 13a bc 3b3 + 13a bc 3c 3 + 13a bc

(b) + + ≥ 6.
(b + c)3 (c + a)3 (a + b)3
a3 b3 c3 3
(a) + + + a b + bc + ca ≥ (a2 + b2 + c 2 );
b+c c+a a+b 2
2a2 + bc 2b2 + ca 2c 2 + a b 9(a2 + b2 + c 2 )

(b) + + ≥ .
b+c c+a a+b 2(a + b + c)
a(b + c) b(c + a) c(a + b)
L
+ 2 + 2 ≥ 2.
b2+ bc + c 2 c + ca + a 2 a + a b + b2
D
.M
A
a(b + c) b(c + a) c(a + b) Y a − b 2
+ + ≥2+4 PI .
b2 + bc + c 2 c 2 + ca + a2 a2 + a b + b2 a+b
M
LY
a b − bc + ca bc − ca + a b ca − a b + bc 3
+ + ≥ .
O
b +c
2 2 c +a
2 2 a +b
2 2 2
.M
1.71. Let a, b, c be nonnegative real numbers, no two of which are zero. If k > −2, then
W
X a b + (k − 1)bc + ca 3(k + 1)
≥ .
b2 + k bc + c2 k+2
W
W
X 3bc − a(b + c) 3
≤ .
b2 + k bc + c2 k+2
ab + 1 bc + 1 ca + 1 4
+ 2 + 2 ≥ .
a +b
2 2 b +c 2 c +a 2 3
12 Vasile Cîrtoaje
5a b + 1 5bc + 1 5ca + 1
+ + ≥ 2.
(a + b)2 (b + c)2 (c + a)2
a2 − bc b2 − ca c2 − a b
+ + ≥ 0.
2b2 − 3bc + 2c 2 2c 2 − 3ca + 2a2 2a2 − 3a b + 2b2
L
.M
2a2 − bc 2b2 − ca 2c 2 − a b
D
+ + ≥ 3.
b2 − bc + c 2 c 2 − ca + a2 a2 − a b + b2
A
PI
M
a2 b2 c2
LY
+ + ≥ 1.
2b2 − bc + 2c 2 2c 2 − ca + 2a2 2a2 − a b + 2b2
O
.M
1 1 1 9
+ 2 + 2 ≥ .
W
4b2 − bc + 4c 2 4c − ca + 4a 2 4a − a b + 4b 2 7(a + b2 + c 2 )
2
W
W
2a2 + bc 2b2 + ca 2c 2 + a b 9
+ 2 + 2 ≥ .
b +c
2 2 c +a 2 a +b 2 2
2a2 + 3bc 2b2 + 3ca 2c 2 + 3a b

+ + ≥ 5.
b2 + bc + c 2 c 2 + ca + a2 a2 + a b + b2
2a2 + 5bc 2b2 + 5ca 2c 2 + 5a b 21

+ + ≥ .
(b + c)2 (c + a)2 (a + b)2 4
X 2a2 + (2k + 1)bc 3(2k + 3)
≥ .
b2 + k bc + c2 k+2
L
.M
3bc − 2a2
D
X 3
≤ .
b + k bc + c
2 2 k+2
A
PI
1.84. If a, b, c are nonnegative real numbers, no two of which are zero, then
M
a2 + 16bc b2 + 16ca c 2 + 16a b
LY
+ 2 + 2 ≥ 10.
b2 + c 2 c + a2 a + b2
O
.M
a2 + 128bc b2 + 128ca c 2 + 128a b

+ + ≥ 46.
W
b2 + c 2 c 2 + a2 a2 + b2
W
W
a2 + 64bc b2 + 64ca c 2 + 64a b

+ + ≥ 18.
(b + c)2 (c + a)2 (a + b)2
1.87. Let a, b, c be nonnegative real numbers, no two of which are zero. If k ≥ −1, then
X a2 (b + c) + ka bc
≥ a + b + c.
b2 + k bc + c 2
14 Vasile Cîrtoaje
−3
1.88. Let a, b, c be nonnegative real numbers, no two of which are zero. If k ≥ ,
2
then
X a3 + (k + 1)a bc
≥ a + b + c.
b2 + k bc + c 2
1.89. Let a, b, c be nonnegative real numbers, no two of which are zero. If k > 0, then
2a k − b k − c k 2b k − c k − a k 2c k − a k − b k
+ 2 + 2 ≥ 0.
b2 − bc + c 2 c − ca + a2 a − a b + b2
L
1.90. If a, b, c are the lengths of the sides of a triangle, then
.M
b+c−a c+a−b a+b−c 2(a + b + c)
(a) + + ≥ ;
b2 − bc + c 2 c 2 − ca + a2 a2 − a b + b2 a2 + b2 + c 2
D
a2 − 2bc b2 − 2ca c 2 − 2a b
A
(b) + + ≤ 0.
b2 − bc + c 2 c 2 − ca + a2 a2 − a b + b2
PI
M
LY
a2 b2 c2 1
+ + ≤ .
5a + (b + c)
2 2 5b + (c + a)
2 2 5c + (a + b)
2 2 3
O
.M
b2 + c 2 − a2 c 2 + a2 − b2 a2 + b2 − c 2 1
+ + ≥ .
W
2a + (b + c)
2 2 2b + (c + a)
2 2 2c + (a + b)
2 2 2
W
1.93. Let a, b, c be positive real numbers. If k > 0, then

W
3a2 − 2bc 3b2 − 2ca 3c 2 − 2a b 3

+ + ≤ .
ka + (b − c)
2 2 k b + (c − a)
2 2 kc + (a − b)
2 2 k
p
1.94. Let a, b, c be nonnegative real numbers, no two of which are zero. If k ≥ 3 + 7,
then
a b c 9
(a) + 2 + 2 ≥ ;
a2 + k bc b + kca c + ka b (1 + k)(a + b + c)
1 1 1 9
(b) + 2 + 2 ≥ .
ka2 + bc k b + ca kc + a b (k + 1)(a b + bc + ca)
1 1 1 6
+ 2 + 2 ≥ 2 .
2a2 + bc 2b + ca 2c + a b a + b + c + a b + bc + ca
2 2
1 1 1 1
+ + ≥ .
22a2 + 5bc 22b + 5ca 22c + 5a b
2 2 (a + b + c)2
L
.M
D
1 1 1 8
+ + ≥ .
2a2 + bc 2b2 + ca 2c 2 + a b (a + b + c)2
A
PI
M
LY
1 1 1 12
+ + ≥ .
a2 + bc b2 + ca c 2 + a b (a + b + c)2
O
.M
1 1 1 1 2
W
(a) + 2 + 2 ≥ 2 + ;
a2 + 2bc b + 2ca c + 2a b a +b +c
2 2 a b + bc + ca
W
a(b + c) b(c + a) c(a + b) a b + bc + ca

(b) + 2 + 2 ≥1+ 2 .
a + 2bc b + 2ca c + 2a b a + b2 + c 2
W
a b c a+b+c
(a) + 2 + 2 ≤ ;
a(b + c) b(c + a) c(a + b) a2 + b2 + c 2

(b) + + ≤ 1 + .
16 Vasile Cîrtoaje
a b c a+b+c
(a) + 2 + 2 ≥ 2 ;
2a2 + bc 2b + ca 2c + a b a + b2 + c 2
b+c c+a a+b 6
(b) + 2 + 2 ≥ .
2a + bc 2b + ca 2c + a b
2 a+b+c
a(b + c) b(c + a) c(a + b) (a + b + c)2

+ +
L
≥ .
a2 + bc b2 + ca c2 + a b a2 + b2 + c 2
.M
1.103. Let a, b, c be nonnegative real numbers, no two of which are zero. If k > 0, then
D
p p p p
A
b2 + c 2 + 3bc c 2 + a2 + 3ca a2 + b2 + 3a b 3(2 + 3)
+ + ≥ .
a2 + k bc b2 + kca PI
c 2 + ka b 1+k
M
LY
1 1 1 8 6
+ 2 + 2 + 2 ≥ .
O
a2 +b 2 b +c 2 c +a 2 a +b +c
2 2 a b + bc + ca
.M

W
a(b + c) b(c + a) c(a + b)

+ 2 + 2 ≤ 2.
a + 2bc b + 2ca c + 2a b
2
W
W
1.106. If a, b, c are real numbers, then
a2 − bc b2 − ca c2 − a b
+ + ≥ 0.
2a2 + b2 + c 2 2b2 + c 2 + a2 2c 2 + a2 + b2
3a2 − bc 3b2 − ca 3c 2 − a b 3
+ + ≤ .
2a + b + c
2 2 2 2b + c + a
2 2 2 2c + a + b
2 2 2 2
(b + c)2 (c + a)2 (a + b)2

+ + ≥ 2.
4a2 + b2 + c 2 4b2 + c 2 + a2 4c 2 + a2 + b2

P 1 3
(a) ≤ ;
11a2 + 2b + 2c
2 2 5(a b + bc + ca)
1 1 1
+
P
L
(b) ≤ .
4a2 +b +c
2 2 2(a + b + c ) a b + bc + ca
2 2 2
.M
1.110. If a, b, c are nonnegative real numbers such that a b + bc + ca = 3, then
D
p p p
A
a b c 3
+ + ≥ .
b+c c+a a+b 2 PI
M
1.111. If a, b, c are nonnegative real numbers such that a b + bc + ca ≥ 3, then
LY
1 1 1 1 1 1
+ + ≥ + + .
2+a 2+ b 2+c 1+ b+c 1+c+a 1+a+ b
O
.M
a2 − bc b2 − ca c2 − a b
W
(a) + + ≤ 0;
3a2 + b2 + c 2 3b2 + c 2 + a2 3c 2 + a2 + b2
W
a4 − b2 c 2 b4 − c 2 a2 c 4 − a2 b2
(b) + + ≤ 0.
3a4 + b4 + c 4 3b4 + c 4 + a4 3c 4 + a4 + b4
W

bc ca ab 1
+ 2 + 2 ≥ .
4a2 +b +c
2 2 4b + c + a
2 2 4c + a + b
2 2 2

1 1 1 9
+ 2 + 2 ≤ .
b2 +c 2 c +a 2 a +b 2 2(a b + bc + ca)
18 Vasile Cîrtoaje

a + b b + c c + a

(a) a − b + b − c + c − a > 5;

a + b2 b2 + c 2 c 2 + a2
2
(b) a2 − b2 + b2 − c 2 + c 2 − a2 ≥ 3.


b+c c+a a+b

a b c
L
+ + +3≥6 + + .
a b c b+c c+a a+b
D
.M
A
X 3a(b + c) − 2bc 3
≥ .
(b + c)(2a + b + c) 2 PI
M
LY
X a(b + c) − 2bc
≥ 0.
(b + c)(3a + b + c)
O
.M
1.119. Let a, b, c be positive real numbers such that a2 + b2 + c 2 ≥ 3. Prove that
a5 − a2 b5 − b2 c5 − c2
W
+ + ≥ 0.
a5 + b2 + c 2 b5 + c 2 + a2 c 5 + a2 + b2
W
W
1.120. Let a, b, c be positive real numbers such that a2 + b2 + c 2 = a3 + b3 + c 3 . Prove

that
a2 b2 c2 3
+ + ≥ .
b+c c+a a+b 2
1.121. If a, b, c ∈ [0, 1], then
a b c
(a) + + ≤ 1;
bc + 2 ca + 2 a b + 2
ab bc ca
(b) + + ≤ 1.
2bc + 1 2ca + 1 2a b + 1
1 1 1

5(1 − a b − bc − ca) + + + 9 ≥ 0.
1 − a b 1 − bc 1 − ca
2 − a2 2 − b2 2 − c2
+ + ≤ 3.
2 − bc 2 − ca 2 − a b
L
.M
3 + 5a2 3 + 5b2 3 + 5c 2
D
+ + ≥ 12.
3 − bc 3 − ca 3 − ab
A
PI
1.125. Let a, b, c be nonnegative real numbers such that a + b + c = 2. If
M
−1 7
≤m≤ ,
LY
7 8
then
O
a2 + m b2 + m c2 + m 3(4 + 9m)
+ + ≥ .
3 − 2bc 3 − 2ca 3 − 2a b 19
.M

W
47 − 7a2 47 − 7b2 47 − 7c 2
W
+ + ≥ 60.
1 + bc 1 + ca 1 + ab
W
26 − 7a2 26 − 7b2 26 − 7c 2 57
+ + ≤ .
1 + bc 1 + ca 1 + ab 2
1.128. Let a, b, c be nonnegative real numbers, no all are zero. Prove that
X 5a(b + c) − 6bc
≤ 3.
a2 + b2 + c 2 + bc
20 Vasile Cîrtoaje
1.129. Let a, b, c be nonnegative real numbers, no two of which are zero, and let
a2 + b2 + c 2
x= .
a b + bc + ca
Prove that
a b c 1 1
(a) + + + ≥x+ ;
b+c c+a a+b 2 x
4

a b c
(b) 6 + + ≥ 5x + ;
b+c c+a a+b x
L
.M
3 1 1

a b c
(c) + + − ≥ x− .
b+c c+a a+b 2 3 x
D
A
1
+ 2
1
+ 2
1
≤
PI 9
.
+ 7(b + c ) b + 7(c + a ) c + 7(a + b ) 5(a + b + c)2
M
a2 2 2 2 2 2 2
LY

O
bc ca ab 3
+ +
.M
≤ .
3a2 + b2 + c 2 3b2 + c 2 + a2 3c 2 + a2 + b2 5
W
1.132. If a, b, c are real numbers such that a + b + c = 3, then

W
1 1 1 3
(a) + + ≤ ;
W
2+ b +c
2 2 2+c +a
2 2 2+a + b
2 2 4
1 1 1 1
(b) + + ≤ .
8 + 5(b + c ) 8 + 5(c + a ) 8 + 5(a + b ) 6
2 2 2 2 2 2
(a + b)(a + c) (b + c)(b + a) (c + a)(c + b) 4

+ 2 + 2 ≤ .
a + 4(b + c ) b + 4(c + a ) c + 4(a + b ) 3
2 2 2 2 2 2 2
X 1 1
≤ .
(b + c)(7a + b + c) 2(a b + bc + ca)
X 1 9
≤ .
b2 + c2 + 4a(b + c) 10(a b + bc + ca)
L
1.136. If a, b, c are nonnegative real numbers such that a + b + c = 3, then
.M
1 1 1 9
+ + ≤ .
D
3 − a b 3 − bc 3 − ca 2(a b + bc + ca)
A
PI
M
bc ca ab 3
+ 2 + 2 ≤ .
a2 +a+6 b + b+6 c +c+6 8
LY
O

.M
1 1 1 1
+ + ≥ .
8a2 − 2bc + 21 8b2 − 2ca + 21 8c 2 − 2a b + 21 9
W
1.139. Let a, b, c be real numbers, no two of which are zero. Prove that
W
a2 + bc b2 + ca c 2 + a b (a + b + c)2
W
(a) + + ≥ ;
b2 + c 2 c 2 + a2 a2 + b2 a2 + b2 + c 2
a2 + 3bc b2 + 3ca c 2 + 3a b 6(a b + bc + ca)

(b) + 2 + 2 ≥ .
b2 + c 2 c + a2 a + b2 a2 + b2 + c 2
1.140. Let a, b, c be real numbers such that a b + bc + ca ≥ 0 and no two of which are
zero. Prove that
a(b + c) b(c + a) c(a + b) 3
+ 2 + 2 ≥ .
b +c
2 2 c +a 2 a +b 2 10
22 Vasile Cîrtoaje
1.141. If a, b, c are positive real numbers such that a bc > 1, then
1 1 4
+ ≥ .
a + b + c − 3 a bc − 1 a b + bc + ca − 3
X (4b2 − ac)(4c 2 − a b) 27
≤ a bc.
b+c 2
L
.M
1.143. Let a, b, c be nonnegative real numbers, no two of which are zero, such that
a + b + c = 3.
D
A
Prove that
a b c 2
+ +
3a + bc 3b + ca 3c + a b
PI
≥ .
3
M
LY
1.144. Let a, b, c be positive real numbers such that
1 1 1

O
(a + b + c) + + = 10.
a b c
.M
Prove that
19 a b c 5
≤ + + ≤ .
12 b+c c+a a+b 3
W
W
1.145. Let a, b, c be nonnegative real numbers, no two of which are zero, such that
W
a + b + c = 3. Prove that
9 a b c
< + + ≤ 1.
10 2a + bc 2b + ca 2c + a b
a3 b3 c3 a3 + b3 + c 3
+ + ≤ .
2a2 + bc 2b2 + ca 2c 2 + a b a2 + b2 + c 2
a3 b3 c3 a+b+c
+ + ≥ .
4a2 + bc 4b2 + ca 4c 2 + a b 5
1 1 1 3
+ + ≥ .
(2 + a)2 (2 + b)2 (2 + c)2 6 + a b + bc + ca
L
.M
1 1 1 3
+ + ≥ .
D
1 + 3a 1 + 3b 1 + 3c 3 + a bc
A
PI
1.150. Let a, b, c be real numbers, no two of which are zero. If 1 ≤ k ≤ 3, then
M
2a b 2bc 2ca

k+ 2 k+ 2 k+ 2 ≥ (k − 1)(k2 − 1).
a + b2 b + c2 c + a2
LY
O
1.151. If a, b, c are non-zero and distinct real numbers, then

.M
1 1 1 1 1 1 1 1 1

+ + +3 + + ≥4 + + .
a2 b2 c 2 (a − b)2 (b − c)2 (c − a)2 a b bc ca
W
W
1.152. Let a, b, c be positive real numbers, and let
a b b c c a
W
A= + + k, B= + + k, C= + + k,
b a c b a b
where −2 < k ≤ 4. Prove that
1 1 1 1 4
+ + ≤ + .
A B C k + 2 A + B + C − (k + 2)
1 1 1 1 1 1
+ 2 + 2 ≥ 2 + 2 + 2 .
b2 + bc + c 2 c + ca + a 2 a + ab + b 2 2a + bc 2b + ca 2c + a b
24 Vasile Cîrtoaje

1 1 1 1 1 1
+ + ≥ 2 + 2 + 2 .
2a b + 1 2bc + 1 2ca + 1 a +2 b +2 c +2

1 1 1 1 1 1
+ + ≥ 2 + 2 + 2 .
a b + 2 bc + 2 ca + 2 a +2 b +2 c +2
L
.M
a b + bc + ca (a − b)2 (b − c)2 (c − a)2
(a) + ≤ 1;
a2 + b2 + c 2 (a2 + b2 )(b2 + c 2 )(c 2 + a2 )
D
A
a b + bc + ca (a − b)2 (b − c)2 (c − a)2
(b) + ≤ 1.
a2 + b2 + c 2 PI
(a2 − a b + b2 )(b2 − bc + c 2 )(c 2 − ca + a2 )
M
LY
1 1 1 45
+ 2 + 2 ≥ .
+b b +c c +a 8(a + b + c ) + 2(a b + bc + ca)
O
a2 2 2 2 2 2 2
.M
1.158. If a, b, c are real numbers, no two of which are zero, then

W
a2 − 7bc b2 − 7ca c 2 − 7a b 9(a b + bc + ca)

+ 2 + 2 + ≥ 0.
b2 + c 2 a + b2 a + b2 a2 + b2 + c 2
W
W
1.159. If a, b, c are real numbers such that a bc 6= 0, then
(b + c)2 (c + a)2 (a + b)2 10(a + b + c)2

+ + ≥ 2 + .
a2 b2 c2 3(a2 + b2 + c 2 )
a2 − 4bc b2 − 4ca c 2 − 4a b 9(a b + bc + ca) 9

+ 2 + 2 + ≥ .
b2 + c 2 a + b2 a + b2 a2 + b2 + c 2 2
a2 + b2 + c 2 9(a − b)2 (b − c)2 (c − a)2

≥1+ .
a b + bc + ca (a + b)2 (b + c)2 (c + a)2
a2 + b2 + c 2 p (a − b)2 (b − c)2 (c − a)2

≥ 1 + (1 + 2)2 2 .
a b + bc + ca (a + b2 )(b2 + c 2 )(c 2 + a2 )
L
.M
2 2 2 5 5 5
+ + ≥ + +
D
.
a+b b+c c+a 3a + b + c 3b + c + a 3c + a + b
A
PI
M
8a2 + 3bc 8b2 + 3ca 8c 2 + 3a b
(a) + + ≥ 11;
b2 + bc + c 2 c 2 + ca + a2 a2 + a b + b2
LY
8a2 − 5bc 8b2 − 5ca 8c 2 − 5a b

(b) + + ≥ 9.
O
b2 − bc + c 2 c 2 − ca + a2 a2 − a b + b2
.M

W
4a2 + bc 4b2 + ca 4c 2 + a b
+ + ≥ 1.
4b2 + 7bc + 4c 2 4c 2 + 7ca + 4a2 4a2 + 7a b + 4b2
W
W
1.166. If a, b, c are real numbers, no two of which are equal, then

1 1 1 27
+ + ≥ .
(a − b)2 (b − c)2 (c − a)2 4(a + b + c − a b − bc − ca)
2 2 2

1 1 1 14
+ + ≥ .
a2 − a b + b2 b2 − bc + c 2 c 2 − ca + a2 3(a2 + b2 + c 2 )
26 Vasile Cîrtoaje
1.168. Let a, b, c be real numbers such that a b + bc + ca ≥ 0 and no two of which are
zero. Prove that
a b c 3
(a) + + ≥ ;
b+c c+a a+b 2
(b) i f a b ≤ 0, then
a b c
+ + ≥ 2.
b+c c+a a+b
L
a b + bc + ca
.M
a b c
+ + ≥ .
7a + b + c 7b + c + a 7c + a + b (a + b + c)2
D
A
a2
+
b2
+
c2 1
≥ .
PI
4a + 5bc 4b + 5ca 4c + 5a b
2 2 2 3
M
LY

O
1 1 1 3
+ + ≥ .
7a2 + b2 + c 2 7b2 + c 2 + a2 7c 2 + a2 + b2 (a + b + c)2
.M
W
1.172. Let a, b, c be the lengths of the sides of a triangle. If k > −2, then
X a(b + c) + (k + 1)bc 3(k + 3)
W
≤ .
b2 + k bc + c2 k+2
W
1.173. Let a, b, c be the lengths of the sides of a triangle. If k > −2, then
X 2a2 + (4k + 9)bc 3(4k + 11)
≤ .
b2 + k bc + c2 k+2
1.174. If a ≥ b ≥ c ≥ d such that a bcd = 1, then

1 1 1 3
+ + ≥ p .
1+a 1+ b 1+c 3
1 + a bc
1.175. Let a, b, c, d be positive real numbers such that a bcd = 1. Prove that
X 1
≤ 1.
1 + a b + bc + ca
1 1 1 1
+ + + ≥ 1.
(1 + a)2 (1 + b)2 (1 + c)2 (1 + d)2
L
.M
1
1.177. Let a, b, c, d 6= be positive real numbers such that a bcd = 1. Prove that
3
D
1 1 1 1
+ + + ≥ 1.
A
(3a − 1)2 (3b − 1)2 (3c − 1)2 (3d − 1)2
PI
M
LY
1 1 1 1
+ + + ≥ 1.
1 + a + a2 + a3 1 + b + b2 + b3 1 + c + c 2 + c 3 1 + d + d 2 + d 3
O
.M
1 1 1 1
W
+ + + ≥ 1.
1 + a + 2a 2 1 + b + 2b 2 1 + c + 2c 2 1 + d + 2d 2
W
W
1 1 1 1 9 25
+ + + + ≥ .
a b c d a+b+c+d 4
1.181. If a, b, c, d are real numbers such that a + b + c + d = 0, then
(a − 1)2 (b − 1)2 (c − 1)2 (d − 1)2

+ + 2 + ≤ 4.
3a2 + 1 3b2 + 1 3c + 1 3d 2 + 1
28 Vasile Cîrtoaje
1.182. If a, b, c, d ≥ −5 such that a + b + c + d = 4, then
1−a 1− b 1−c 1−d

+ + + ≥ 0.
(1 + a)2 (1 + b)2 (1 + c)2 (1 + d)2
1.183. Let a1 , a2 , . . . , an be positive real numbers such that a1 + a2 + · · · + an = n. Prove

that X 1 1
≤ .
(n + 1)a1 + a2 + · · · + an2
2 2 2
L
.M
1.184. Let a1 , a2 , . . . , an be real numbers such that a1 + a2 + · · · + an = 0. Prove that
(a1 + 1)2 (a2 + 1)2 (an + 1)2
D
n
+ + ··· + ≥ .
a12 + n − 1 a22 + n − 1 an + n − 1
2 n−1
A
PI
1.185. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. Prove that
M
LY
1 1 1
+ + ··· + ≥ 1.
1 + (n − 1)a1 1 + (n − 1)a2 1 + (n − 1)an
O
.M
1 1 1
+ + ··· + ≥ 1.
W
1 − a1 + na12 1 − a2 + na22 1 − an + nan2

W
W
1.187. Let a1 , a2 , . . . , an be positive real numbers such that
k(n − k − 1)
a1 , a2 , . . . , an ≥ , k>1
kn − k − 1
and
a1 a2 · · · an = 1.
Prove that
1 1 1 n
+ + ··· + ≤ .
a1 + k a2 + k an + k 1+k
1.188. Let a1 , a2 , . . . , an be positive real numbers such that
a1 ≥ 1 ≥ a2 ≥ · · · ≥ an , a1 a2 · · · an = 1.
Prove that
1 − a1 1 − a2 1 − an
+ + ··· + ≥ 0.
3 + a12 3 + a22 3 + an2
1.189. If a1 , a2 , . . . , an ≥ 0, then
1 1 1 n
L
+ + ··· + ≥ .
1 + na1 1 + na2 1 + nan n + a1 a2 · · · an
.M
D
1.190. If a1 , a2 , . . . , an are positive real numbers, then
A
b1 b2 bn a1 a2 an
+
a1 a2
+ ··· +
an
≥ +
b1 b2
+ ··· + ,
bn
PI
M
where
1 X
bi = i = 1, 2, · · · , n.
LY
aj,
n − 1 j6=i
O
.M
1.191. If a1 , a2 , . . . , an are positive real numbers such that
1 1 1
a1 + a2 + · · · + an = + + ··· + ,
W
a1 a2 an
then
W
1 1 1
(a) + + ··· + ≥ 1;
1 + (n − 1)a1 1 + (n − 1)a2 1 + (n − 1)an
W
1 1 1
(b) + + ··· + ≤ 1.
n − 1 + a1 n − 1 + a2 n − 1 + an
30 Vasile Cîrtoaje
L
D
.M
A
PI
M
LY
O
.M
W
W
W
1.2 Solutions
P 1.1. If a, b are nonnegative real numbers, then
1 1 1
+ ≥ .
(1 + a)2 (1 + b)2 1 + ab
(Vasile Cîrtoaje, 1994)
First Solution. Use the Cauchy-Schwarz inequality as follows:
L
1 1 1 (b + a)2 1
.M
+ − ≥ −
(1 + a)2 (1 + b)2 1 + ab b (1 + a) + a (1 + b)
2 2 2 2 1 + ab
a b[a2 + b2 − 2(a + b) + 2]
=
D
(1 + a b)[b2 (1 + a)2 + a2 (1 + b)2 ]
A
a b[(a − 1)2 + (b − 1)2 ]
= ≥ 0.
PI
(1 + a b)[b2 (1 + a)2 + a2 (1 + b)2 ]
The equality holds for a = b = 1.

M
Second Solution. By the Cauchy-Schwarz inequality, we have
LY
1 1

(a + b) a + ≥ (a + 1)2 , (a + b) + b ≥ (1 + b)2 ,
O
b a
.M
hence
1 1 1 1 1
+ ≥ + = .
W
(1 + a)2 (1 + b)2 (a + b)(a + 1/b) (a + b)(1/a + b) 1 + a b

W
Third Solution. The desired inequality follows from the identity

W
1 1 1 a b(a − b)2 + (1 − a b)2

+ − = .
(1 + a)2 (1 + b)2 1 + a b (1 + a)2 (1 + b)2 (1 + a b)
P 1.2. If a, b, c are nonnegative real numbers, then
a2 − bc b2 − ca c2 − a b
+ + ≥ 0.
3a + b + c 3b + c + a 3c + a + b
32 Vasile Cîrtoaje
Solution. We use the SOS method. Without loss of generality, assume that a ≥ b ≥ c.
We have
X a2 − bc X (a − b)(a + c) + (a − c)(a + b)
2 =
3a + b + c 3a + b + c
X (a − b)(a + c) X (b − a)(b + c)
= +
3a + b + c 3b + c + a
X (a − b)2 (a + b − c)
=
(3a + b + c)(3b + c + a)
Since a + b − c ≥ 0, it suffices to show that
L
(b − c)2 (b + c − a)(3a + b + c) + (c − a)2 (c + a − b)(3b + c + a) ≥ 0;
.M
that is,
D
(a − c)2 (c + a − b)(3b + c + a) ≥ (b − c)2 (a − b − c)(3a + b + c).
A
PI
This inequality is trivial for a ≤ b + c. Otherwise, we can get it by multiplying the
obvious inequalities
M
c + a − b ≥ a − b − c,
LY
b2 (a − c)2 ≥ a2 (b − c)2 ,
a(3b + c + a) ≥ b(3a + b + c),
O
a ≥ b.
.M
The equality holds for a = b = c, and for a = 0 and b = c (or any cyclic permutation).
W
W
P 1.3. If a, b, c are positive real numbers, then

W
4a2 − b2 − c 2 4b2 − c 2 − a2 4c 2 − a2 − b2
+ + ≤ 3.
a(b + c) b(c + a) c(a + b)
Solution. We use the SOS method. Write the inequality as follows:

X 4a2 − b2 − c 2

1− ≥ 0,
a(b + c)
X b2 + c 2 − 4a2 + a(b + c)
≥ 0,
a(b + c)
X (b2 − a2 ) + a(b − a) + (c 2 − a2 ) + a(c − a)

≥ 0,
a(b + c)
X (b − a)(2a + b) + (c − a)(2a + c)
≥ 0,
a(b + c)
X (b − a)(2a + b) X (a − b)(2b + a)
+ ≥ 0,
a(b + c) b(c + a)
X
c(a + b)(a − b)2 (bc + ca − a b) ≥ 0.
Without loss of generality, assume that a ≥ b ≥ c. Since ca + a b − bc > 0, it suffices to
L
show that
.M
b(c + a)(c − a)2 (a b + bc − ca) + c(a + b)(a − b)2 (bc + ca − a b) ≥ 0,
D
that is,
A
b(c + a)(a − c)2 (a b + bc − ca) ≥ c(a + b)(a − b)2 (a b − bc − ca).
PI
For the nontrivial case a b − bc − ca > 0, this inequality follows by multiplying the
M
inequalities
LY
a b + bc − ca > a b − bc − ca,
O
(a − c)2 ≥ (a − b)2 ,
b(c + a) ≥ c(a + b).

.M
The equality holds for a = b = c

W
W
W
P 1.4. Let a, b, c be nonnegative real numbers, no two of which are zero. Prove that
1 1 1 3
(a) + 2 + 2 ≥ ;
a2 + bc b + ca c + a b a b + bc + ca
1 1 1 2
(b) + 2 + 2 ≥ .
2a2 + bc 2b + ca 2c + a b a b + bc + ca
1 1 1 2
(c) + 2 + 2 > .

34 Vasile Cîrtoaje
Solution. (a) Since

a b + bc + ca a(b + c − a)
=1+ ,
a + bc
2 a2 + bc
we can write the inequality as
a(b + c − a) b(c + a − b) c(a + b − c)

+ + ≥ 0.
a2 + bc b2 + ca c2 + a b
Without loss of generality, assume that a = min{a, b, c}. Since b + c − a > 0, it suffices
to show that
b(c + a − b) c(a + b − c)
+ ≥ 0.
b2 + ca c2 + a b
L
This is equivalent to each of the following inequalities
.M
(b2 + c 2 )a2 − (b + c)(b2 − 3bc + c 2 )a + bc(b − c)2 ≥ 0,
D
(b − c)2 a2 − (b + c)(b − c)2 a + bc(b − c)2 + a bc(2a + b + c) ≥ 0,
A
(b − c)2 (a − b)(a − c) + a bc(2a + b + c) ≥ 0.
PI
The last inequality is obviously true. The equality holds for a = 0 and b = c (or any
M
cyclic permutation thereof).
LY
(b) According to the identities
2a2 + bc = a(2a − b − c) + a b + bc + ca,

O
.M
2b2 + ca = b(2b − c − a) + a b + bc + ca,

2c 2 + a b = c(2c − a − b) + a b + bc + ca,
W

W
1 1 1
+ + ≥ 2,
1+ x 1+ y 1+z
W
where
a(2a − b − c) b(2b − c − a) c(2c − a − b)
x= , y= , z= .
a b + bc + ca a b + bc + ca a b + bc + ca
1
Without loss of generality, assume that a = min{a, b, c}. Since x ≤ 0 and ≥ 1, it
1+ x
suffices to show that
1 1
+ ≥ 1.
1+ y 1+z
1 ≥ yz,
(a b + bc + ca)2 ≥ bc(2b − c − a)(2c − a − b),
a2 (b2 + bc + c 2 ) + 3a bc(b + c) + 2bc(b − c)2 ≥ 0.

The last inequality is obviously true. The equality holds for a = 0 and b = c (or any
cyclic permutation thereof).
(c) According to the identities
a2 + 2bc = (a − b)(a − c) + a b + bc + ca,
b2 + 2ca = (b − c)(b − a) + a b + bc + ca,
L
c 2 + 2a b = (c − a)(c − b) + a b + bc + ca,
.M
D
1 1 1
+ + > 2,
1+ x 1+ y 1+z
A
where
(a − b)(a − c)
PI
(b − c)(b − a) (c − a)(c − b)
x= , y= , z= .
M
LY
Since
x y + yz + z x = 0
O
and
−(a − b)2 (b − c)2 (c − a)2
x yz =
.M
≤ 0,
(a b + bc + ca)3
we have
1 1 1 1 − 2x yz
W
+ + −2= > 0.
1+ x 1+ y 1+z (1 + x)(1 + y)(1 + z)
W
W
a(b + c) b(c + a) c(a + b)

+ 2 + 2 ≥ 2.
a2 + bc b + ca c + ab
(Pham Kim Hung, 2006)
Solution. Without loss of generality, assume that a ≥ b ≥ c and write the inequality as
b(c + a) (a − b)(a − c) (a − c)(b − c)

≥ + .
b2 + ca a2 + bc c2 + a b
36 Vasile Cîrtoaje
Since
(a − b)(a − c) (a − b)a a−b
≤ 2 ≤
a + bc
2 a + bc a
and
(a − c)(b − c) a(b − c) b−c
≤ 2 ≤ ,
c + ab
2 c + ab b
it suffices to show that
b(c + a) a−b b−c
≥ + .
b2 + ca a b
This inequality is equivalent to
L
b2 (a − b)2 − 2a bc(a − b) + a2 c 2 + a b2 c ≥ 0,
.M
(a b − b2 − ac)2 + a b2 c ≥ 0.
The equality holds for for a = 0 and b = c (or any cyclic permutation).
D
A
PI
M
a2 b2 c2 a b c
LY
+ + ≥ + + .
b +c
2 2 c +a
2 2 a +b
2 2 b+c c+a a+b
O
Solution. We have
.M
X a2 a
X a b(a − b) + ac(a − c)
− =
b2 + c 2 b + c (b2 + c 2 )(b + c)
W
a b(a − b) ba(b − a)
W
X X
= +
(b + c )(b + c)
2 2 (c + a2 )(c + a)
2
W
X a b(a − b)2
= (a2 + b2 + c 2 + a b + bc + ca) ≥ 0.
(b2 + c 2 )(c 2 + a2 )(b + c)(c + a)
The equality holds for a = b = c, and also for a = 0 and b = c (or any cyclic permuta-
tion).
P 1.7. Let a, b, c be positive real numbers. Prove that
1 1 1 a b c
+ + ≥ 2 + 2 + 2 .
b+c c+a a+b a + bc b + ca c + a b
First Solution. Without loss of generality, assume that a = min{a, b, c}. Since
X 1 X a X 1 a

− = −
b+c a2 + bc b + c a2 + bc
X (a − b)(a − c)
=
(b + c)(a2 + bc)
and (a − b)(a − c) ≥ 0, it suffices to show that
(b − c)(b − a) (c − a)(c − b)
+ ≥ 0.
(c + a)(b + ca) (a + b)(c 2 + a b)
2
L
.M
(b − c)[(b2 − a2 )(c 2 + a b) + (a2 − c 2 )(b2 + ca)] ≥ 0,
D
a(b − c)2 (b2 + c 2 − a2 + a b + bc + ca) ≥ 0,
A
and is clearly true for a = min{a, b, c}. The equality holds for a = b = c.
Second Solution. Since
PI
M
1 X 1 1
X
X b c
= + = a + ,
b+c (b + c)2 (b + c)2 (a + b)2 (a + c)2
LY

O
X 1 1 1

a + − ≥ 0.
.M
(a + b)2 (a + c)2 a2 + bc
This is true, since

W
1 1 1 bc(b − c)2 + (a2 − bc)2

+ − = ≥ 0.
W
(a + b)2 (a + c)2 a2 + bc (a + b)2 (a + c)2 (a2 + bc)

W
We can also prove this inequality using the Cauchy-Schwarz inequality, as follows
1 1 1 (c + b)2 1
+ − ≥ −
(a + b)2 (a + c)2 a2 + bc c 2 (a + b)2 + b2 (a + c)2 a2 + bc
bc[2a2 − 2a(b + c) + b2 + c 2 ]
= 2
(a + bc)[c 2 (a + b)2 + b2 (a + c)2 ]
bc[(2a − b − c)2 + (b − c)2 ]
= ≥ 0.
2(a2 + bc)[c 2 (a + b)2 + b2 (a + c)2 ]
38 Vasile Cîrtoaje
1 1 1 2a 2b 2c
+ + ≥ 2 + 2 + 2 .
b+c c+a a+b 3a + bc 3b + ca 3c + a b
Solution. Since
1 2a X 1 2a
X X
− = −
b+c 3a2 + bc b + c 3a2 + bc
X (a − b)(a − c) + a(2a − b − c)
= ,
(b + c)(3a2 + bc)
L
.M
it suffices to show that X (a − b)(a − c)
≥0
(b + c)(3a2 + bc)
D
and
A
X a(2a − b − c)
≥ 0.
(b + c)(3a2 + bc) PI
In order to prove the first inequality, assume that a = min{a, b, c}. Since
M
(a − b)(a − c) ≥ 0,
LY
it is enough to show that

O
(b − c)(b − a) (c − a)(c − b)
+ ≥ 0.
.M
(c + a)(3b2 + ca) (a + b)(3c 2 + a b)

This is equivalent to the obvious inequality
W
a(b − c)2 (b2 + c 2 − a2 + 3a b + bc + 3ca) ≥ 0.

W
The second inequality is also true, since

W
X a(2a − b − c) X a(a − b) + a(a − c)

=
(b + c)(3a2 + bc) (b + c)(3a2 + bc)
X a(a − b) X b(b − a)
= +
(b + c)(3a + bc)
2 (c + a)(3b2 + ca)

X a b
= (a − b) −
(b + c)(3a2 + bc) (c + a)(3b2 + ca)
X c(a − b)2 [(a − b)2 + c(a + b)]
= ≥ 0.
(b + c)(c + a)(3a2 + bc)(3b2 + ca)
The equality holds for a = b = c.
a b c 13 2(a b + bc + ca)
(a) + + ≥ − ;
b+c c+a a+b 6 3(a2 + b2 + c 2 )
3 p a b + bc + ca

a b c
(b) + + − ≥ ( 3 − 1) 1 − 2 .
b+c c+a a+b 2 a + b2 + c 2
Solution. (a) We use the SOS method. Rewrite the inequality as
3 2 a b + bc + ca

a b c
+ + − ≥ 1− 2 .
L
b+c c+a a+b 2 3 a + b2 + c 2
.M
Since
X a 1 (a − b) + (a − c)
X
D
− =
b+c 2 2(b + c)
A
X a−b X b−a
= +
2(b + c)
X a−b 1
PI
2(c + a)
1

= −
M
2 b+c c+a
(a − b)2
LY
X
=
2(b + c)(c + a)
O
and
2 a b + bc + ca (a − b)2
X
1− 2 = ,
.M
3 a + b2 + c 2 3(a2 + b2 + c 2 )
the inequality can be restated as
W
1 1
X
(a − b)2 − ≥ 0.
2(b + c)(c + a) 3(a2 + b2 + c 2 )
W
This is true, since

W
3(a2 + b2 + c 2 ) − 2(b + c)(c + a) = (a + b − c)2 + 2(a − b)2 ≥ 0.

(b) Let
p = a + b + c, q = a b + bc + ca, r = a bc.
We have
X a X a X 1
= +1 −3= p −3
b+c b+c b+c
p(p2 + q)
= − 3.
pq − r
40 Vasile Cîrtoaje
According to P 3.57-(a) in Volume 1, for fixed p and q, the product r is minimal when
a = 0 or b = c. Therefore, it suffices to prove the inequality for a = 0 and for b = c = 1.
Case 1: a = 0. The original inequality can be written as
b c 3 p

bc
+ − ≥ ( 3 − 1) 1 − 2 .
c b 2 b + c2
It suffices to show that

b c 3 bc
+ − ≥1− 2 .
c b 2 b + c2
L
Denoting
.M
b2 + c 2
t= , t ≥ 2,
bc
D
this inequality becomes
3 1
A
t− ≥1− ,
2 t
(t − 2)(2t − 1) ≥ 0.
PI
M
Case 2: b = c = 1. The original inequality becomes as follows:
LY
2 3 p 2a + 1

a
+ − ≥ ( 3 − 1) 1 − 2 ,
2 a+1 2 a +2
O
p
(a − 1)2 ( 3 − 1)(a − 1)2
≥ ,
.M
2(a + 1) a2 + 2
p
(a − 1)2 (a − 3 + 1)2 ≥ 0.
W
a
The equality holds for a = b = c, and for p = b = c (or any cyclic permutation).
3−1
W
W

2
1 1 1 a+b+c

+ + ≤ .
First Solution. Write the inequality as
(a + b + c)2 X a b + bc + ca
−3≥ − 1 ,
a b + bc + ca a2 + 2bc
(a − b)2 + (b − c)2 + (a − b)(b − c) X (a − b)(a − c)

+ ≥ 0.
Without loss of generality, assume that a ≥ b ≥ c. Since (a − b)(a − c) ≥ 0 and (c −
a)(c − b) ≥ 0, it suffices to show that
(a b + bc + ca)(a − b)(b − c)
(a − b)2 + (b − c)2 + (a − b)(b − c) − ≥ 0.
b2 + 2ca
(a − b)2 (b − c)2
(a − b)2 + (b − c)2 − ≥ 0,
b2 + 2ca
or
L
c(a − b)2 (2a + 2b − c)
(b − c)2 + ≥ 0,
.M
b2 + 2ca
which is clearly true. The equality holds for a = b = c.
Second Solution. Assume that a ≥ b ≥ c and write the desired inequality as
D
(a + b + c)2 X a b + bc + ca
A

−3≥ −1 ,
1 X
PI
X (a − b)(a − c)
(a − b)(a − c) + ≥ 0,
M
a b + bc + ca
X
LY
1+ (a − b)(a − c) ≥ 0.
a2 + 2bc
Since (c − a)(c − b) ≥ 0 and a − b ≥ 0, it suffices to prove that
O
a b + bc + ca a b + bc + ca

1+ (a − c) + 1 + (c − b) ≥ 0.
.M
a2 + 2bc b2 + 2ca
Write this inequality as
W
a−c c−b

a − b + (a b + bc + ca) 2 + ≥ 0,
a + 2bc b2 + 2ca
W
(a b + bc + ca)(3ac + 3bc − a b − 2c 2

(a − b) 1 + ≥ 0.
(a2 + 2bc)(b2 + 2ca)
W
Since a − b ≥ 0 and 2ac + 3bc − 2c 2 > 0, it is enough to show that

(a b + bc + ca)(ac − a b)
1+ ≥ 0.
(a2 + 2bc)(b2 + 2ca)
We have
(a b + bc + ca)(ac − a b) (a b + bc + ca)(ac − a b)
1+ ≥1+
(a2 + 2bc)(b2 + 2ca) a2 (b2 + ca)
(a + b)c + (a2 − b2 )c
2
= > 0.
a(b2 + ca)
42 Vasile Cîrtoaje
a2 (b + c) b2 (c + a) c 2 (a + b)
+ 2 + 2 ≥ a + b + c.
b2 + c 2 c + a2 a + b2
(Darij Grinberg, 2004)
First Solution. We use the SOS method. We have

X a2 (b + c) X X a2 (b + c)
− a= −a
b2 + c 2 b2 + c 2
X a b(a − b) + ac(a − c)
=
L
b2 + c 2
.M
X a b(a − b) X ba(b − a)
= +
b2 + c 2 c 2 + a2
D
X a b(a + b)(a − b)2
= ≥ 0.
(b2 + c 2 )(c 2 + a2 )
A
PI
tion).
M
Second Solution. By virtue of the Cauchy-Schwarz inequality, we have
LY
[ a2 (b + c)]2
P
X a2 (b + c)
≥P .
b2 + c 2 a2 (b + c)(b2 + c 2 )
O
Then, it suffices to show that

.M
X X X
[ a2 (b + c)]2 ≥ ( a)[ a2 (b + c)(b2 + c 2 )].
W
Let p = a + b + c and q = a b + bc + ca. Since

W
X
[ a2 (b + c)]2 = (pq − 3a bc)2
W
= p2 q2 − 6a bc pq + 9a2 b2 c 2
and
X X
a2 (b + c)(b2 + c 2 ) = (b + c)[(a2 b2 + b2 c 2 + c 2 a2 ) − b2 c 2 ]
X
= 2p(a2 b2 + b2 c 2 + c 2 a2 ) − b2 c 2 (p − a)
= p(a2 b2 + b2 c 2 + c 2 a2 ) + a bcq = p(q2 − 2a bc p) + a bcq,
the inequality can be written as
a bc(2p3 + 9a bc − 7pq) ≥ 0.
Using Schur’s inequality

p3 + 9a bc − 4pq ≥ 0,
we have
2p3 + 9a bc − 7pq ≥ p(p2 − 3q) ≥ 0.
a2 + b2 b2 + c 2 c 2 + a2 3(a2 + b2 + c 2 )
+ + ≤ .
L
a+b b+c c+a a + b + c)
.M
Solution. We use the SOS method.
D
First Solution. Multiplying by 2(a + b + c), the inequality successively becomes
A
X a 2
1+ (b + c 2 ) ≤ 3(a2 + b2 + c 2 ),
PI
b+c
X a
M
X
(b2 + c 2 ) ≤ a2 ,
b+c
LY
X b2 + c 2

a a− ≥ 0,
b+c
O
X a b(a − b) − ac(c − a)
≥ 0,
.M
b+c
X a b(a − b) X ba(a − b)
− ≥ 0,
b+c c+a
W
X a b(a − b)2
≥ 0.
W
(b + c)(c + a)
W
tion).
Second Solution. Subtracting a + b + c from the both sides, the desired inequality
becomes as follows:
3(a2 + b2 + c 2 ) X a2 + b2 a + b
− (a + b + c) ≥ − ,
a+b+c a+b 2
X (a − b)2 X (a − b)2
≥ ,
a+b+c 2(a + b)
X (a + b − c)(a − b)2
≥ 0.
a+b
44 Vasile Cîrtoaje
Without loss of generality, assume that a ≥ b ≥ c. Since a + b − c ≥ 0, it suffices to

prove that
(a + c − b)(a − c)2 (a − b − c)(b − c)2
≥ .
a+c b+c
a−c b−c
This inequality is true because a + c − b ≥ a − b − c, a − c ≥ b − c and ≥ .
a+c b+c
The last inequality reduces to c(a − b) ≥ 0.
Third Solution. Write the inequality as follows
X 3(a2 + b2 ) a2 + b2

− ≥ 0,
2(a + b + c) a+b
L
.M
X (a2 + b2 )(a + b − 2c)
≥ 0,
a+b
D
X (a2 + b2 )(a − c) X (a2 + b2 )(b − c)
+ ≥ 0,
A
a+b a+b
+
PI
X (a2 + b2 )(a − c) X (b2 + c 2 )(c − a)
≥ 0,
a+b b+c
M
X (a − c)2 (a b + bc + ca − b2 )
≥ 0.
LY
(a + b)(b + c)
It suffices to prove that
O
X (a − c)2 (a b + bc − ca − b2 )
.M
≥ 0.
(a + b)(b + c)
Since a b + bc − ca − b2 = (a − b)(b − c), this inequality is equivalent to

W
X c−a
(a − b)(b − c)(c − a) ≥ 0,
W
(a + b)(b + c)
W
which is true because X c−a

= 0.
(a + b)(b + c)
1 1 1 9
+ 2 + 2 ≥ .
a2 + ab + b 2 b + bc + c 2 c + ca + a 2 (a + b + c)2

First Solution. Due to homogeneity, we may assume that a + b + c = 1. Let q = a b +

bc + ca. Since
b2 + bc + c 2 = (a + b + c)2 − a(a + b + c) − (a b + bc + ca) = 1 − a − q,

X 1
≥ 9,
1−a−q
9q3 − 6q2 − 3q + 1 + 9a bc ≥ 0.
L
From Schur’s inequality
.M
(a + b + c)3 + 9a bc ≥ 4(a + b + c)(a b + bc + ca),
D
we get 1 + 9a bc − 4q ≥ 0. Therefore,
A
9q3 − 6q2 − 3q + 1 + 9a bc = (1 + 9a bc − 4q) + q(3q − 1)2 ≥ 0.

PI
M
Second Solution. Multiplying by a2 + b2 +c 2 +a b+ bc +ca, the inequality can be written
LY
as X a 9(a b + bc + ca)
(a + b + c) + ≥ 6.
b + bc + c
2 2 (a + b + c)2
O
By the Cauchy-Schwarz inequality, we have

.M
X a (a + b + c)2 a+b+c
≥ = .
b + bc + c a(b + bc + c ) a b + bc + ca
P
2 2 2 2
W

W
(a + b + c)2 9(a b + bc + ca)

+ ≥ 6.
W
a b + bc + ca (a + b + c)2
This follows immediately from the AM-GM inequality.
a2 b2 c2 1
+ + ≤ .
(2a + b)(2a + c) (2b + c)(2b + a) (2c + a)(2c + b) 3
(Tigran Sloyan, 2005)

46 Vasile Cîrtoaje
First Solution. The inequality is equivalent to each of the inequalities

X a2

a
− ≤ 0,
(2a + b)(2a + c) 3(a + b + c)
X a(a − b)(a − c)
≥ 0.
(2a + b)(2a + c)
Due to symmetry, we may consider that a ≥ b ≥ c. Since c(c − a)(c − b) ≥ 0, it suffices
to prove that
a(a − b)(a − c) b(b − c)(b − a)
+ ≥ 0.
(2a + b)(2a + c) (2b + c)(2b + a)
L
.M
(a − b)2 [(a + b)(2a b − c 2 ) + c(a2 + b2 + 5a b)] ≥ 0.
D
A
tion).
PI
Second Solution (by Vo Quoc Ba Can). Apply the Cauchy-Schwarz inequality in the
following manner
M
9a2 (2a + a)2 2a a2
LY
= ≤ + .
(2a + b)(2a + c) 2a(a + b + c) + (2a2 + bc) a + b + c 2a2 + bc
O
Then,
X 9a2 X a2
≤2+ ,
.M
(2a + b)(2a + c) 2a2 + bc

and from the known inequality
W
X a2
≤ 1,
2a2 + bc
W
the conclusion follows. The last inequality is equivalent to

W
X bc
≥ 1,
2a2 + bc
and can be obtained using the Cauchy-Schwarz inequality, as follows
( bc)2
P
X bc
≥P = 1.
2a2 + bc bc(2a2 + bc)
Remark. From the inequality in P 1.14 and Hölder’s inequality
a2
X
X Æ 2
a(2a + b)(2a + c) ≥ (a + b + c)3 ,
(2a + b)(2a + c)
we get the following result:

• If a, b, c are nonnegative real numbers such that a + b + c = 3, then
Æ Æ Æ
a(2a + b)(2a + c) + b(2b + c)(2b + a) + c(2c + a)(2c + bc) ≥ 9,
3 3
with equality for a = b = c = 1, and for (a, b, c) = (0, , ) or any cyclic permutation.
2 2
L
a 1
.M
P
(a) ≤ ;
(2a + b)(2a + c) a+b+c
D
P a3 1
(b) ≤ .
(2a2 + b2 )(2a2 + c 2 ) a+b+c
A
PI (Vasile Cîrtoaje, 2005)
M
Solution. (a) Write the inequality as
a(a + b + c)
LY
X1
− ≥ 0,
3 (2a + b)(2a + c)
O
X (a − b)(a − c)
≥ 0.
(2a + b)(2a + c)
.M
Assume that a ≥ b ≥ c. Since (a − b)(a − c) ≥ 0, it suffices to prove that
(b − c)(b − a) (a − c)(b − c)
W
+ ≥ 0.
(2b + c)(2b + a) (2c + a)(2c + b)
W
Since b − c ≥ 0 and a − c ≥ a − b ≥ 0, it is enough to show that

W
1 1
≥ .
(2c + a)(2c + b) (2b + c)(2b + a)
(b − c)(a + 4b + 4c) ≥ 0.

(b) We obtain the desired inequality by summing the inequalities
a3 a
≥ ,
(2a + b )(2a + c ) (a + b + c)2
2 2 2 2
48 Vasile Cîrtoaje
b3 b
≥ ,
(2b + c )(2b + a ) (a + b + c)2
2 2 2 2
c3 c
≥ ,
(2c + a )(2c + b ) (a + b + c)2
2 2 2 2
which are consequences of the Cauchy-Schwarz inequality. For example, from
(a2 + a2 + b2 )(c 2 + a2 + a2 ) ≥ (ac + a2 + ba)2 ,
the first inequality follows. The equality holds for a = b = c.
L
.M
D
X 1 1 2
≥ + .
A
(a + 2b)(a + 2c) (a + b + c)2 3(a b + bc + ca)
PI
Solution. Write the inequality as follows
M
X 1 1 2 2

LY
− ≥ − ,
(a + 2b)(a + 2c) (a + b + c) 2 3(a b + bc + ca) (a + b + c)2
O
X (b − c)2 X (b − c)2
≥ ,
(a + 2b)(a + 2c) 3(a b + bc + ca)
.M
X b−c
(a − b)(b − c)(c − a) ≥ 0.
(a + 2b)(a + 2c)
W
Since
W
b−c X b−c b−c

X
= −
(a + 2b)(a + 2c) (a + 2b)(a + 2c) 3(a b + bc + ca)
W
(a − b)(b − c)(c − a) X 1
= ,
3(a b + bc + ca) (a + 2b)(a + 2c)
the desired inequality is equivalent to the obvious inequality

X 1
(a − b)2 (b − c)2 (c − a)2 ≥ 0.
(a + 2b)(a + 2c)
The equality holds for a = b, or b = c, or c = a.

1 1 1 4
(a) + + ≥ ;
(a − b)2 (b − c)2 (c − a)2 a b + bc + ca
1 1 1 3
(b) + + ≥ ;
a2 − a b + b2 b2 − bc + c 2 c 2 − ca + a2 a b + bc + ca
1 1 1 5
(c) + 2 + 2 ≥ .
a2 +b 2 b +c 2 c +a 2 2(a b + bc + ca)
L
Solution. Let
.M
Ek (a, b, c) = + 2 + 2 .
a − ka b + b
2 2 b − k bc + c 2 c − kca + a2
D
We will prove that
A
Ek (a, b, c) ≥ αk ,
where 
5 − 2k
PI
, 0≤k≤1
M

2−k

αk = .
 2 + k, 1≤k≤2
LY
To show this, assume that a ≤ b ≤ c and prove that

O
Ek (a, b, c) ≥ Ek (0, b, c) ≥ αk .
.M
For the nontrivial case a > 0, the left inequality is true because
W
Ek (a, b, c) − Ek (0, b, c)
=
W
a
b2 + (1 + k)bc − ac b+c c 2 + (1 + k)bc − a b
= + +
W
b(a2 − ka b + b2 ) b2 − k bc + c 2 c(c 2 − kca + a2 )

bc − ac b+c bc − a b
> + 2 + > 0.
b(a − ka b + b ) b − k bc + c
2 2 2 c(c − kca + a2 )
2
In order to prove the right inequality Ek (0, b, c) ≥ αk , where
bc b c
Ek (0, b, c) = + + ,
b2 − k bc + c 2 c b
by virtue of the AM-GM inequality, we have
bc b2 − k bc + c 2
Ek (0, b, c) = + + k ≥ 2 + k.
b2 − k bc + c 2 bc
50 Vasile Cîrtoaje
Thus, it remains to consider the case 0 ≤ k ≤ 1. We have
bc b2 − k bc + c 2
Ek (0, b, c) = +
b2 − k bc + c 2 (2 − k)2 bc
1

b c k
+ 1− + +
(2 − k) 2 c b (2 − k)2
2 1 5 − 2k

k
≥ +2 1− + = .
2−k (2 − k)2 (2 − k)2 2−k
b c
For 1 ≤ k ≤ 2, the equality holds when a = 0 and + = 1 + k (or any cyclic
c b
permutation). For 0 ≤ k ≤ 1, the equality holds when a = 0 and b = c (or any cyclic
L
permutation).
D
.M
A
P 1.18. Let a, b, c be positive real numbers, no two of which are zero. Prove that
+ +
PI
(a2 + b2 )(a2 + c 2 ) (b2 + c 2 )(b2 + a2 ) (c 2 + a2 )(c 2 + b2 )
≥ a2 + b2 + c 2 .
(a + b)(a + c) (b + c)(b + a) (c + a)(c + b)
M
LY
Solution. Using the identity

O
(a2 + b2 )(a2 + c 2 ) = a2 (a2 + b2 + c 2 ) + b2 c 2 ,

.M
we can write the inequality as follows
b2 c 2 a2
X X
W
2 2 2
≥ (a + b + c ) 1 − ,
(a + b)(a + c) (a + b)(a + c)
W
X
b2 c 2 (b + c) ≥ 2a bc(a2 + b2 + c 2 ),
W
X X
a3 (b2 + c 2 ) ≥ 2 a3 bc,
X
a3 (b − c)2 ≥ 0.
Since the last form is obvious, the proof is completed. The equality holds for a = b = c.
P 1.19. Let a, b, c be positive real numbers such that a + b + c = 3. Prove that

1 1 1
+ 2 + 2 ≤ 1.
First Solution. By virtue of the Cauchy-Schwarz inequality, we have
(a2 + b + c)(1 + b + c) ≥ (a + b + c)2 .
Therefore,
X 1 X 1+ b+c 3 + 2(a + b + c)
≤ = = 1.
a +b+c
2 (a + b + c)2 (a + b + c)2
The equality occurs for a = b = c = 1.
Second Solution. Rewrite the inequality as
L
1 1 1
+ +
.M
≤ 1.
a2 − a + 3 b2 − b + 3 c 2 − c + 3
We see that the equality holds for a = b = c = 1. Thus, if there exists a real number k
D
such that
A
1 1

≤k+ −k a
a2 − a + 3 3 PI
for all a ∈ [0, 3], then
M
1 X 1 1
X
LY
X
≤ k+ − k a = 3k + −k a = 1.
a2 − a + 3 3 3
O
We have
.M
1 1 (a − 1)[(1 − 3k)a2 + 3ka + 3(1 − 3k)]

k+ −k a− 2 = .
3 a −a+3 3(a2 − a + 3)
W
Setting k = 4/9, we get

W
1 1 (a − 1)2 (3 − a)

k+ −k a− 2 = ≥ 0.
W
3 a −a+3 9(a2 − a + 3)
P 1.20. Let a, b, c be real numbers such that a + b + c = 3. Prove that
a2 − bc b2 − ca c 2 − a b
+ 2 + 2 ≥ 0.
a2 + 3 b +3 c +3

52 Vasile Cîrtoaje
Solution. Apply the SOS method. We have

X a2 − bc X (a − b)(a + c) + (a − c)(a + b)
2 =
a2 + 3 a2 + 3
X (a − b)(a + c) X (b − a)(b + c)
= +
a2 + 3 b2 + 3
a+c b+c
X
= (a − b) 2 − 2
a +3 b +3
X (a − b)2
= (3 − a b − bc − ca) ≥ 0.
(a2 + 3)(b2 + 3)
L
Thus, it suffices to show that 3 − a b − bc − ca ≥ 0. This follows immediately from the
.M
known inequality (a + b + c)2 ≥ 3(a b + bc + ca), which is equivalent to (a − b)2 + (b −
c)2 + (c − a)2 ≥ 0. The equality holds for a = b = c = 1.
D
A
PI
P 1.21. Let a, b, c be nonnegative real numbers such that a + b + c = 3. Prove that
M
1 − bc 1 − ca 1 − a b
+ + ≥ 0.
5 + 2a 5 + 2b 5 + 2c
LY
Solution. Since
O
9(1 − bc) = (a + b + c)2 − 9bc,

.M

X a2 + b2 + c 2 + 2a(b + c) − 7bc
≥ 0.
W
5 + 2a
W
From
(a − b)(a + k b + mc) + (a − c)(a + kc + mb) =
W
= 2a2 − k(b2 + c 2 ) + (k + m − 1)a(b + c) − 2mbc,

choosing k = −2 and m = 7, we get
(a − b)(a − 2b + 7c) + (a − c)(a − 2c + 7b) = 2[a2 + b2 + c 2 + 2a(b + c) − 7bc].
Therefore, the desired inequality becomes as follows:

X (a − b)(a − 2b + 7c) X (a − c)(a − 2c + 7b)
+ ≥ 0,
5 + 2a 5 + 2a
X (a − b)(a − 2b + 7c) X (b − a)(b − 2a + 7c)
+ ≥ 0,
5 + 2a 5 + 2b
X
(a − b)(5 + 2c)[(5 + 2b)(a − 2b + 7c) − (5 + 2a)(b − 2a + 7c)] ≥ 0,
X
(a − b)2 (5 + 2c)(15 + 4a + 4b − 14c) ≥ 0,
X
(a − b)2 (5 + 2c)(a + b − c) ≥ 0.
Without loss of generality, assume that a ≥ b ≥ c. Clearly, it suffices to show that
(a − c)2 (5 + 2b)(a + c − b) ≥ (b − c)2 (5 + 2a)(a − b − c).
Since a − c ≥ b − c ≥ 0 and a + c − b ≥ a − b − c, it suffices to prove that
(a − c)(5 + 2b) ≥ (b − c)(5 + 2a).
L
.M
Indeed,
(a − c)(5 + 2b) − (b − c)(5 + 2a) = (a − b)(5 + 2c) ≥ 0.
D
The equality holds for a = b = c = 1, and for c = 0 and a = b = 3/2 (or any cyclic
A
permutation).
PI
M
LY
1 1 1 3
+ 2 + 2 ≤ .
O
a2 + b +2 b +c +2 c +a +2 4
2 2 2
.M
Solution. Since
1 1 a2 + b2
W
= − ,
a2 + b2 + 2 2 a2 + b2 + 2
W
we write the inequality as

W
a2 + b2 b2 + c 2 c 2 + a2 3
+ + ≥ .
a + b +2 b +c +2 c +a +2 2
2 2 2 2 2 2

Pp Pp
( a2 + b2 )2 2 a2 + 2 (a2 + b2 )(a2 + c 2 )
P
X a2 + b2
≥ =
a2 + b2 + 2 (a2 + b2 + 2) 2 a2 + 6
P P
2 a + 2 (a + bc) 3 a + 9 3
P 2 P 2 P 2
≥ = P = .
2 a2 + 6 2 a2 + 6 2
P
The equality holds for a = b = c = 1

54 Vasile Cîrtoaje
1 1 1 1
+ 2 + 2 ≤ .
4a2 +b +c
2 2 4b + c + a
2 2 4c + a + b
2 2 2
Solution. According to the Cauchy-Schwarz inequality, we have
9 (a + b + c)2
=
4a2 + b2 + c 2 2a2 + (a2 + b2 ) + (a2 + c 2 )
1 b2 c2
L
≤ + 2 + .
2 a + b2 a2 + c 2
.M
Therefore,
D
b2 c2

X 9 3 X
≤ + +
A
4a2 + b2 + c 2 2 a2 + b2 a2 + c 2
3 X
= +
2

b2
+
a2
a2 + b2 b2 + a2
PI

3
2
9
= +3= .
2
M
The equality holds for a = b = c = 1.
LY
O
.M
bc ca ab
+ 2 + 2 ≤ 1.
W
a2 +1 b +1 c +1
W
Solution. Let p = a + b + c and q = a b + bc + ca, q ≤ p2 /3 = 4/3. If one of a, b, c is

W
zero, then the inequality is true. Otherwise, write the inequality as

X 1 1
≤ ,
+ 1)
a(a2 a bc
X1 a

1
− 2 ≤ ,
a a +1 a bc
X a 1 1 1 1
≥ + + − ,
a +1
2 a b c a bc
X a q−1
≥ ,
a +1
2 a bc
Using the inequality

2
≥ 2 − a,
a2 +1
which is equivalent to a(a − 1)2 ≥ 0, we get
X a X a(2 − a) X a(b + c)
≥ = = q.
a2 + 1 2 2
Therefore, it suffices to prove that
1 + a bcq ≥ q.
L
.M
Since
D
a4 + b4 + c 4 = (a2 + b2 + c 2 )2 − 2(a2 b2 + b2 c 2 + c 2 a2 )
A
= (p2 − 2q)2 − 2(q2 − 2a bc p) = p4 − 4p2 q + 2q2 + 4a bc p
by Schur’s inequality of degree four

PI
M
a4 + b4 + c 4 + 2a bc(a + b + c) ≥ (a b + bc + ca)(a2 + b2 + c 2 ),
LY
we get
O
(p2 − q)(4q − p2 )
a bc ≥ ,
6p
.M
(4 − q)(q − 1)
a bc ≥ .
3
W
Thus
W
q(4 − q)(q − 1) (3 − q)(q − 1)2

1 + a bcq − q ≥ 1 + −q = ≥ 0.
3 3
W
The equality holds if a = 0 and b = c = 1 (or any cyclic permutation).
bc ca ab 1
+ + ≤ .
a+1 b+1 c+1 4

56 Vasile Cîrtoaje
First Solution. We have

X bc X bc
=
a+1 (a + b) + (c + a)
1X 1 1

≤ bc +
4 a+b c+a
1 X bc 1 X bc
= +
4 a+b 4 c+a
1 X bc 1 X ca
= +
4 a+b 4 a+b
1 X bc + ca 1X 1
= = c= .
L
4 a+b 4 4
.M
The equality holds for a = b = c = 1/3, and for a = 0 and b = c = 1/2 (or any cyclic
permutation).
D
Second Solution. It is easy to check that the inequality is true if one of a, b, c is zero.
A
Otherwise, write the inequality as
1
+
1
+
1
≤
PI
1
.
a(a + 1) b(b + 1) c(c + 1) 4a bc
M
LY
Since
1 1 1
= − ,
a(a + 1) a a+1
O
we can write the required inequality as

.M
1 1 1 1 1 1 1
+ + ≥ + + − .
a+1 b+1 c+1 a b c 4a bc
W
In virtue of the Cauchy-Schwarz inequality,

W
1 1 1 9 9
+ + ≥ = .
a + 1 b + 1 c + 1 (a + 1) + (b + 1) + (c + 1) 4
W
9 1 1 1 1
≥ + + − .
4 a b c 4a bc
This is equivalent to Schur’s inequality
(a + b + c)3 + 9a bc ≥ 4(a + b + c)(a b + bc + ca).


1 1 1 3
+ + ≤ .
a(2a + 1) b(2b + 1) c(2c + 1) 11a bc
2 2 2

Solution. Since
1 1 2a
= − 2 ,
a(2a + 1)
2 a 2a + 1
X 2a 1 1 1 3
≥ + + − .
2a2 + 1 a b c 11a bc
L
.M
2( a)2
P
X 2a 2
≥P = .
D
2a + 1
2 a(2a + 1) 2(a + b + c 3 ) + 1
2 3 3
A
Therefore, it suffices to show that
2(a3
2
+ +c )+1
b33
≥
11q − 3
11a bc
,
PI
M
1 1
where q = a b + bc + ca, q ≤ (a + b + c)2 = . Since
LY
3 3
a3 + b3 + c 3 = 3a bc + (a + b + c)3 − 3(a + b + c)(a b + bc + ca) = 3a bc + 1 − 3q,
O
we need to prove that

.M
22a bc ≥ (11q − 3)(6a bc + 3 − 6q),

or, equivalently,
W
2(20 − 33q)a bc ≥ 3(11q − 3)(1 − 2q).

W

W
we get
9a bc ≥ 4q − 1.
Thus,
2(20 − 33q)a bc − 3(11q − 3)(1 − 2q) ≥
2(20 − 33q)(4q − 1)
≥ − 3(11q − 3)(1 − 2q)
9
330q2 − 233q + 41 (1 − 3q)(41 − 110q)
= = ≥ 0.
9 9
This completes the proof. The equality holds for a = b = c = 1/3.
58 Vasile Cîrtoaje
1 1 1
+ 3 + 3 ≤ 1.
Solution. Write the inequality in the form
1 1 1
+ + ≤ 1.
a3 − a + 3 b3 − b + 3 c 3 − c + 3
Assume that a ≥ b ≥ c. There are two cases to consider.
L
.M
Case 1: c ≤ b ≤ a ≤ 2. The desired inequality follows by adding the inequalities
1 5 − 2a 1 5 − 2b 1 5 − 2c
D
≤ , 3 ≤ , 3 ≤ .
a3 −a+3 9 b − b+3 9 c −c+3 9
A
Indeed, we have
1 5 − 2a
PI
(a − 1)2 (a − 2)(2a + 3)
=
M
− ≤ 0.
a3 − a + 3 9 9(a3 − a + 3)
LY
Case 2: a > 2. From a + b + c = 3, we get b + c < 1. Since

O
X 1 1 1 1 1 1 1
< 3 + + < + +
.M
,
a3 −a+3 a −a+3 3− b 3−c 9 3− b 3−c
it suffices to prove that

W
1 1 8
+ ≤ .
3− b 3−c 9
W
We have
1 1 8 −3 − 15(1 − b − c) − 8bc
W
+ − = < 0.
3− b 3−c 9 9(3 − b)(3 − c)
a2 b2 c2
+ + ≥ 1.
1 + b3 + c 3 1 + c 3 + a3 1 + a3 + b3
Solution. Using the Cauchy-Schwarz inequality, we have
( a2 )2
P
X a2
≥P ,
1 + b3 + c 3 a2 (1 + b3 + c 3 )
and it remains to show that

X
(a2 + b2 + c 2 )2 ≥ (a2 + b2 + c 2 ) + a2 b2 (a + b).
Let p = a + b + c and q = a b + bc + ca, q ≤ 3. Since a2 + b2 + c 2 = 9 − 2q and

X X X
a2 b2 (a + b) = a2 b2 (3 − c) = 3 a2 b2 − qa bc = 3q2 − (q + 18)a bc,
L
.M
the desired inequality can be written as
q2 − 34q + 72 + (q + 18)a bc ≥ 0.
D
A
This inequality is clearly true for q ≤ 2. Consider further that 2 < q ≤ 3. Since
PI
a4 + b4 + c 4 = (a2 + b2 + c 2 )2 − 2(a2 b2 + b2 c 2 + c 2 a2 )
= (p2 − 2q)2 − 2(q2 − 2a bc p) = p4 − 4p2 q + 2q2 + 4a bc p
M
LY
by Schur’s inequality of degree four
a4 + b4 + c 4 + 2a bc(a + b + c) ≥ (a b + bc + ca)(a2 + b2 + c 2 ),
O
we get
.M
(p2 − q)(4q − p2 ) (9 − q)(4q − 9)

a bc ≥ = .
6p 18
W
Therefore
(q + 18)(9 − q)(4q − 9)
W
q2 − 34q + 72 + (q + 18)a bc ≥ q2 − 34q + 72 +

18
W
(3 − q)(4q2 + 21q − 54)

= ≥ 0.
18

1 1 1 3
+ + ≤ .
6 − a b 6 − bc 6 − ca 5
60 Vasile Cîrtoaje
Solution. Rewrite the inequality as
108 − 48(a b + bc + ca) + 13a bc(a + b + c) − 3a2 b2 c 2 ≥ 0,
4[9 − 4(a b + bc + ca) + 3a bc] + a bc(1 − a bc) ≥ 0.

By the AM-GM inequality,
3
a+b+c

1= ≥ a bc.
3
Consequently, it suffices to show that
9 − 4(a b + bc + ca) + 3a bc ≥ 0.
L
We see that the homogeneous form of this inequality is just Schur’s inequality of third
.M
degree
(a + b + c)3 + 9a bc ≥ 4(a + b + c)(a b + bc + ca).
D
The equality holds for a = b = c = 1, as well as for a = 0 and b = c = 3/2 (or any cyclic
A
permutation).
PI
M
LY
1 1 1 1
+ 2 + 2 ≤ .
2a2 + 7 2b + 7 2c + 7 3
O

.M
Solution. Assume that a = max{a, b, c} and prove that

1
W
E(a, b, c) ≤ E(a, s, s) ≤ ,
3
W
where
b+c
s= , 0 ≤ s ≤ 1,
W
2
and
1 1 1
E(a, b, c) = + + .
2a2 + 7 2b2 + 7 2c 2 + 7
We have
1 1 1 1

E(a, s, s) − E(a, b, c) = − + −
2s2 + 7 2b2 + 7 2s2 + 7 2c 2 + 7
1 (b − c)(b + s) (c − b)(c + s)

= 2 +
2s + 7 2b2 + 7 2c 2 + 7
(b − c) (7 − 4s − 2bc)
2 2
= .
(2s2 + 7)(2b2 + 7)(2c 2 + 7)
Since bc ≤ s2 ≤ 1, it follows that 7 − 4s2 − 2bc > 0, and hence E(a, s, s) ≥ E(a, b, c).
Also,
1 1 4(s − 1)2 (2s − 1)2
− E(a, s, s) = − E(3 − 2s, s, s) = ≥ 0.
3 3 3(2a2 + 7)(2s2 + 7)
permutation).
P 1.31. Let a, b, c be nonnegative real numbers such that a ≥ b ≥ 1 ≥ c and a + b + c = 3.
L
Prove that
1 1 1 3
+ +
.M
≤ .
a2 + 3 b2 + 3 c 2 + 3 4
D
First Solution (by Nguyen Van Quy). Write the inequality as follows:
A
1 3−a 1 3− b 3−c 1

a2 + 3
−
8
+
b2 + 3
−
8
≤ PI 8
− 2
c +3
,
M
(a − 1)3 (b − 1)3 (1 − c)3
+ ≤ .
a2 + 3 b2 + 3 c2 + 3
LY
Indeed, we have
O
(1 − c)3 (a − 1 + b − 1)3 (a − 1)3 + (b − 1)3 (a − 1)3 (b − 1)3

= ≥ ≥ + 2 .
c2 + 3 c2 + 3 c2 + 3 a2 + 3 b +3
.M
The proof is completed. The equality holds for a = b = c = 1.

W
Second Solution. Let d be a positive number such that
c + d = 2.
W
We have
W
a + b = 1 + d, d ≥ a ≥ b ≥ 1.
In addition, we claim that
1 1 1
+ 2 ≤ .
c2 +3 d +3 2
Indeed,
1 1 1 (cd − 1)2
− 2 − 2 = ≥ 0.
2 c + 3 d + 3 2(c 2 + 3)(d 2 + 3)
Thus, it suffices to show that
1 1 1 1
+ ≤ 2 + .
a2 + 3 b2 + 3 d +3 4
62 Vasile Cîrtoaje
Since
1 1 (d − a)(d + a) 1 1 (b − 1)(b + 1)
− 2 = 2 , − 2 = ,
a2 + 3 d + 3 (a + 3)(d 2 + 3) 4 b +3 4(b2 + 3)
d+a b+1
≤ .
(a2 + 3)(d + 3) 4(b2 + 3)
2
We can get this inequality by multiplying the inequalities

d+a a+1
≤ ,
d +3
2 4
L
a+1 b+1
.M
≤ 2 .
a +3
2 b +3
We have
D
a+1 d+a (d − 1)(ad + a + d − 3)
− 2 = ≥ 0,
4 d +3 4(d 2 + 3)
A
b+1 a+1 (a − b)(a b + a + b − 3)
PI
− 2 = ≥ 0.
b +3 a +3
2 (a2 + 3)(b2 + 3)
M
LY

O
1 1 1 3
+ 2 + 2 ≥ .
.M
2a2 + 3 2b + 3 2c + 3 5
W
First Solution (by Nguyen Van Quy). Write the inequality as

W
X1 1

2
− 2 ≤ ,
3 2a + 3 5
W
X a2 3
≤ .
2a + 5 5
2
Using the Cauchy-Schwarz inequality gives

25 25
=
3(2a2 + 3) 6a2 + (a + b + c)2
(2 + 2 + 1)2
=
2(2a2 + bc) + 2a(a + b + c) + a2 + b2 + c 2
22 22 1
≤ + + 2 ,
2(2a + bc) 2a(a + b + c) a + b2 + c 2
2
hence
X 25a2 X 2a2 X 2a X a2
≤ + +
3(2a2 + 3) 2a2 + bc a+b+c a2 + b2 + c 2
X 2a2
= + 3.
2a2 + bc
X a2
≤ 1,
2a2 + bc
L
which is equivalent to
X1 a2

1
.M
− 2 ≥ ,
2 2a + bc 2
bc
D
X
≥ 1.
2a + bc
2
A
Using again the Cauchy-Schwarz inequality, we get
X bc ( bc)2
P PI
≥P = 1.
M
2a2 + bc bc(2a2 + bc)
LY
permutation).
O
Second Solution. First, we can check that the desired inequality becomes an equality
for a = b = c = 1, and for a = 0 and b = c = 3/2. Consider then the inequality
.M
f (x) ≥ 0, where
1
f (x) = − A − B x.
2x + 3
W
We have
−4x
W
f 0 (x) = − B.
(2x 2 + 3)2
W
From the conditions f (1) = 0 and f 0 (1) = 0, we get A = 9/25 and B = −4/25. Also,
from the conditions f (3/2) = 0 and f 0 (3/2) = 0, we get A = 22/75 and B = −8/75.
From these values of A and B, we obtain the identities
1 9 − 4x 2(x − 1)2 (4x − 1)

− = ,
2x 2 + 3 25 25(2x 2 + 3)
1 22 − 8x (2x − 3)2 (4x + 1)

− = ,
2x 2 + 3 75 75(2x 2 + 3)
and the inequalities
1 9 − 4x 1
≥ , x≥ ,
2x 2 + 3 25 4
64 Vasile Cîrtoaje
1 22 − 8x
≥ , x ≥ 0.
2x 2 + 3 75
Without loss of generality, assume that a ≥ b ≥ c.
1
Case 1: a ≥ b ≥ c ≥ . By summing the inequalities
4
1 9 − 4a 1 9 − 4b 1 9 − 4c
≥ , ≥ , ≥ ,
2a2 + 3 25 2b2 + 3 25 2c 2 + 3 25
we get
1 1 1 27 − 4(a + b + c) 3
+ 2 + 2 ≥ = .
2a2+ 3 2b + 3 2c + 3
L
25 5
.M
1
Case 2: a ≥ b ≥ ≥ c. We have
4
D
X 1 22 − 8a 22 − 8b 1
+ + 2
A
≥
2a2 +3 75 75 2c + 3
=
44 − 8(a + b)
75
+ 2
1
2c + 3
=
PI
20 + 8c
75
+ 2
1
2c + 3
.
M
LY
20 + 8c 1 3
+ 2 ≥ ,
75 2c + 3 5
O
which is equivalent to the obvious inequality

.M
c(8c 2 − 25c + 12) ≥ 0.

W
1
W
Case 3: a ≥ ≥ b ≥ c. We have
4
W
X 1 1 1 2 3
> + 2 ≥ > .
2a2 +3 2b2 + 3 2c + 3 1
8 +3 5
P 1.33. Let a, b, c be nonnegative real numbers such that a ≥ 1 ≥ b ≥ c and a + b + c = 3.

Prove that
1 1 1
+ 2 + 2 ≥ 1.
a +2 b +2 c +2
2

First Solution. Let b1 and c1 be positive numbers such that
b + b1 = 2, c + c1 = 2.
We have
b1 + c1 = 1 + a, 1 ≤ b1 ≤ c1 ≤ a.
In addition, we claim that
1 1 2 1 1 2
+ 2 ≥ , + 2 ≥ .
b2 + 2 b1 + 2 3 c2 + 2 c1 + 2 3
L
Indeed,
.M
1 1 2 2b b1 (1 − b b1 ) b b1 (b − b1 )2
+ − = = ≥ 0,
b2 + 2 b12 + 2 3 3(b2 + 2)(b12 + 2) 6(b2 + 2)(b12 + 2)
D
cc1 (c − c1 )2
A
1 1 2
+ − = ≥ 0.
c 2 + 2 c12 + 2 3 6(c 2 + 2)(c12 + 2) PI
Using these inequalities, it suffices to show that
M
1 1 1 1
+ ≥ 2 + 2 .
LY
a2 +2 3 b1 + 2 c1 + 2
O
Since
1 1 (b1 − 1)(b1 + 1) 1 1 (a − c1 )(a + c1 )
.M
− 2 = , − = ,
3 b1 + 2 3(b12 + 2) c12 +2 a2 + 2 (a2 + 2)(c12 + 2)
W

b1 + 1 a + c1
≥ .
W
3(b12 + 2) (a2 + 2)(c12 + 2)

We can get this inequality by multiplying the inequalities
W
b1 + 1 c1 + 1
≥ ,
b12 +2 c12 + 2
c1 + 1 a + c1
≥ 2 .
3 a +2
We have
b1 + 1 c1 + 1 (c1 − b1 )(b1 c1 + b1 + c1 − 2)
− = ≥ 0,
b12 +2 c12 +2 (b12 + 2)(c12 + 2)
c1 + 1 a + c1 (a − 1)(ac1 + a + c1 − 2)
− 2 = ≥ 0.
3 a +2 3(a2 + 2)
66 Vasile Cîrtoaje
The proof is completed. The equality holds for a = b = c = 1, as well as for a = 2, b = 1

and c = 0.
Second Solution. First, we can check that the desired inequality becomes an equality
for a = b = c = 1, and also for a = 2, b = 1, c = 0. Consider then the inequality
f (x) ≥ 0, where
1
f (x) = 2 − A − B x.
x +2
We have
−2x
f 0 (x) = 2 − B.
(x + 2)2
L
From the conditions f (1) = 0 and f 0 (1) = 0, we get A = 5/9 and B = −2/9. Also, from
the conditions f 2) = 0 and f 0 (2) = 0, we get A = 7/18 and B = −1/9. From these
.M
values of A and B, we obtain the identities
D
1 5 − 2x (x − 1)2 (2x − 1)
− = ,
x2 + 2 9(x 2 + 2)
A
9
1
−
7 − 2x
=
PI
(x − 2)2 (2x + 1)
,
x2 + 2 18 18(x 2 + 2)
M
and the inequalities
LY
1 5 − 2x 1
≥ , x≥ ,
x2 + 2 9 2
O
1 7 − 2x
≥ , x ≥ 0.
x +2
2 18
.M
Let us define
1
g(x) = .
x2 + 2
W
p
Notice that for d ∈ (0, 2] and x ∈ [0, d], we have
W
g(d) − g(0)
g(x) ≥ g(0) + x,
W
d
because
g(d) − g(0) x(d − x)(2 − d x)
g(x) − g(0) − x= ≥ 0.
d 2(d 2 + 2)(x 2 + 2)
For d = 1/2 and d = 1, we get the inequalities
1 9 − 2x 1
≥ , 0≤ x ≤ ,
x2 +2 18 2
1 3− x
≥ , 0 ≤ x ≤ 1.
x2 + 2 6
Consider further two cases: c ≥ 1/2 and c ≤ 1/2.
1
Case 1: c ≥ . By summing the inequalities
2
1 5 − 2a 1 5 − 2b 1 5 − 2c
≥ , ≥ , ≥ ,
a2 + 2 9 b2 + 2 9 c2 + 2 9
we get
1 1 1 15 − 2(a + b + c)
+ 2 + 2 ≥ = 1.
a2 +2 b +2 c +2 9
1
Case 2: c ≤ . We have
2
1 7 − 2a
L
≥ ,
a2 +2 18
.M
1 3− b 8 − 2b
≥ ≥ ,
b2 + 2 6 18
D
1 9 − 2c
A
≥ .
c2 + 2 18
Therefore, PI
M
1 1 1 7 − 2a 8 − 2b 9 − 2c
+ 2 + 2 ≥ + + = 1.
a2 +2 b +2 c +2 18 18 18
LY
O
.M
P 1.34. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. Prove that
a+b+c
W
1 1 1 3
+ + ≥ + .
a+b b+c c+a 6 a+b+c
W

W
First Solution. Denoting x = a + b + c, we have
1 1 1 (a + b + c)2 + a b + bc + ca x2 + 3
+ + = = .
a+b b+c c+a (a + b + c)(a b + bc + ca) − a bc 3x − a bc
Then, the inequality becomes
x2 + 3 x 3
≥ + ,
3x − a bc 6 x
or
3(x 3 + 9a bc − 12x) + a bc(x 2 − 9) ≥ 0.
68 Vasile Cîrtoaje
This inequality is true since from (a + b + c)2 ≥ 3(a b + bc + ca), we get x 2 − 9 ≥ 0, and
from Schur’s inequality of degree three

3
we get xp + 9a bc − 12 ≥ 0. The equality holds for a = b = c = 1, and for a = 0 and
b = c = 3 (or any cyclic permutation).
Second Solution. We apply the SOS method. Write the inequality as follows:
1 1 1 a+b+c 3
+ + ≥ + ,
a+b b+c c+a 2(a b + bc + ca) a + b + c
L
(a + b + c)2
.M
1 1 1

2(a + b + c) + + ≥ + 6,
a+b b+c c+a a b + bc + ca
D
1 1 1 (a + b + c)2

[(a + b) + (b + c) + (c + a)] + + −9≥ − 3,
a+b b+c c+a a b + bc + ca
A
X (b − c)2 1 PI X
≥ (b − c)2 ,
(a + b)(c + a) 2(a b + bc + ca)
M
X a b + bc + ca − a2
(b − c)2 ≥ 0.
(a + b)(c + a)
LY
Without loss of generality, assume that a ≥ b ≥ c. Since a b + bc + ca − c 2 ≥ 0, it suffices

O
to show that
.M
a b + bc + ca − a2 a b + bc + ca − b2
(b − c)2 + (c − a)2 ≥ 0.
(a + b)(c + a) (b + c)(a + b)
W
Since a b + bc + ca − b2 ≥ b(a − b) ≥ 0 and (c − a)2 ≥ (b − c)2 , it is enough to prove that

W
a b + bc + ca − a2 a b + bc + ca − b2
(b − c)2 + (b − c)2 ≥ 0.
(a + b)(c + a) (b + c)(a + b)
W
This is true if
a b + bc + ca − a2 a b + bc + ca − b2
+ ≥ 0, .
(a + b)(c + a) (b + c)(a + b)
3 − a2 3 − b2
+ ≥ 0,
3 + a2 3 + b2
Indeed,
3 − a2 3 − b2 2(9 − a2 b2 ) 2c(a + b)(3 + a b)
+ = = ≥ 0.
3+a 2 3+ b 2 (3 + a )(3 + b )
2 2 (3 + a2 )(3 + b2 )
1 1 1 3
+ 2 + 2 ≥ .
a2 +1 b +1 c +1 2
First Solution. After expanding, the inequality can be restated as
a2 + b2 + c 2 + 3 ≥ a2 b2 + b2 c 2 + c 2 a2 + 3a2 b2 c 2 .
From
L
(a + b + c)(a b + bc + ca) − 9a bc = a(b − c)2 + b(c − a)2 + c(a − b)2 ≥ 0,
.M
we get
D
a + b + c ≥ 3a bc.
A
So, it suffices to show that
PI
a2 + b2 + c 2 + 3 ≥ a2 b2 + b2 c 2 + c 2 a2 + a bc(a + b + c).
M
This is equivalent to the homogeneous inequalities
LY
(a b + bc + ca)(a2 + b2 + c 2 ) + (a b + bc + ca)2 ≥ 3(a2 b2 + b2 c 2 + c 2 a2 ) + 3a bc(a + b + c),

O
a b(a2 + b2 ) + bc(b2 + c 2 ) + ca(c 2 + a2 ) ≥ 2(a2 b2 + b2 c 2 + c 2 a2 ),

.M
a b(a − b)2 + bc(b − c)2 + ca(c − a)2 ≥ 0.

p
The equality holds for a = b = c = 1, and for a = 0 and b = c = 3 (or any cyclic
W
permutation).
W
Second Solution. Without loss of generality, assume that a = min{a, b, c}. From a b +
bc + ca = 3, we get bc ≥ 1. Also, from
W
(a + b + c)(a b + bc + ca) − 9a bc = a(b − c)2 + b(c − a)2 + c(a − b)2 ≥ 0,
we get
a + b + c ≥ 3a bc.
The desired inequality follows by summing the inequalities
1 1 2
+ 2 ≥ ,
b2 +1 c +1 bc + 1
1 2 3
+ ≥ .
a2 + 1 bc + 1 2
70 Vasile Cîrtoaje
We have
1 1 2 b(c − b) c(b − c)
+ 2 − = 2 + 2
b2 + 1 c + 1 bc + 1 (b + 1)(bc + 1) (c + 1)(bc + 1)
(b − c)2 (bc − 1)
= 2 ≥0
(b + 1)(c 2 + 1)(bc + 1)
and
1 2 3 a2 − bc + 3 − 3a2 bc a(a + b + c − 3a bc)

+ − = = ≥ 0.
a + 1 bc + 1 2
2 2(a + 1)(bc + 1)
2 2(a2 + 1)(bc + 1)
L
Third Solution. Since
.M
1 a2 1 b2 1 c2
= 1 − , = 1 − , = 1 − ,
a2 + 1 a2 + 1 b2 + 1 b2 + 1 c 2 + 1 c2 + 1
D
A
we can rewrite the inequality as
a2
+
b2
+
c2 3
≤ ,
PI
a +1 b +1 c +1 2
2 2 2
M
or, in the homogeneous form
LY
X a2 1
≤ .
O
3a + a b + bc + ca
2 2
.M
According to the Cauchy-Schwarz inequality, we have
4a2 (a + a)2 a a2
= ≤ + .
W
3a2 + a b + bc + ca a(a + b + c) + (2a2 + bc) a + b + c 2a2 + bc

W
Then,
X 4a2 X a2
≤ 1 + ,
W
3a2 + a b + bc + ca 2a2 + bc
and it suffices to show that
X a2
≤ 1,
2a2 + bc
or, equivalently,
X bc
≥ 1.
2a2 + bc
This follows from the Cauchy-Schwarz inequality as follows:
( bc)2 b c + 2a bc a
P P 2 2 P
X bc
≥P = = 1.
2a2 + bc bc(2a2 + bc) 2a bc a + b2 c 2
P P
Remark. We can write the inequality in P 1.35 in the homogeneous form

1 1 1 3
2
+ 2
+ 2
≥ .
3a 3b 3c 2
1+ 1+ 1+
1 1 1
Substituting a, b, c by , , , respectively, we get
x y z
x y z 3
+ + ≥ .
3 yz 3z x 3x y 2
x+ y+ z+
x + y +z x + y +z x + y +z
L
So, we find the following result.
.M
• If x, y, z are positive real numbers such that x + y + z = 3, then
D
x y z 3
+ + ≥ .
x + yz y + zx z + x y
A
2
PI
M
P 1.36. Let a, b, c be positive real numbers such that a b + bc + ca = 3. Prove that
LY
a2 b2 c2
+ + ≥ 1.
O
a2 + b + c b2 + c + a c 2 + a + b
.M
Solution. We apply the Cauchy-Schwarz inequality in the following way

W
2
a3/2 + b3/2 + c 3/2 a + 2 (a b)3/2
P 3 P
X a2
≥ P = .
a2 + b + c a(a2 + b + c) a3 + 6
P
W
Then, we still have to show that

W
(a b)3/2 + (bc)3/2 + (ca)3/2 ≥ 3.
(a b)3/2 + (a b)3/2 + 1 1 3a b 1
(a b)3/2 = − ≥ − ,
2 2 2 2
and hence X 3X 3
(a b)3/2 ≥ a b − = 3.
2 2
72 Vasile Cîrtoaje
bc + 4 ca + 4 a b + 4 bc + 2 ca + 2 a b + 2
+ 2 + 2 ≤3≤ 2 + + .
a +4 b +4 c +4
2 a + 2 b2 + 2 c 2 + 2
Solution. More general, using the SOS method, we will show that
bc + k ca + k a b + k

(k − 3) 2 + + 2 −3 ≤0
a + k b2 + k c +k
for k > 0. This inequality is equivalent to
L
.M
X a2 − bc
(k − 3) ≥ 0.
a2 + k
D
Since
A
X a2 − bc X (a − b)(a + c) + (a − c)(a + b)
2
a2 + k
=
a2 + k
PI
M
X (a − b)(a + c) X (b − a)(b + c)
= +
a2 + k b2 + k
LY
X (a − b)2
= (k − a b − bc − ca)
(a2 + p)(b2 + p)
O
X (a − b) 2
= (k − 3) ,
.M
(a + p)(b2 + p)
2
we have
X a2 − bc (a − b)2
W
X
2(k − 3) = (k − 3)2 ≥ 0.
a2 + k (a2 + k)(b2 + k)
W
Equality in both inequalities holds for a = b = c = 1.

W
p
P 1.38. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. If k ≥ 2 + 3,
then
1 1 1 3
+ + ≤ .
a+k b+k c+k 1+k
Solution. Let us denote p = a + b + c, p ≥ 3. By expanding, the inequality becomes
k(k − 2)p + 3a bc ≥ 3(k − 1)2 .

Since this inequality is true for p ≥ 3(k − 1)2 /(k2 − 2k), consider further that
3(k − 1)2
p≤ .
k(k − 2)
(a + b + c)3 + 9a bc ≥ 4(a b + bc + ca)(a + b + c),
we get 9a bc ≥ 12p − p3 . Therefore, it suffices to prove that
3k(k − 2)p + 12p − p3 ≥ 9(k − 1)2 ,
L
.M
or, equivalently,
(p − 3)[(3(k − 1)2 − p2 − 3p] ≥ 0.
D
Thus, it remains to prove that
A
3(k − 1)2 − p2 − 3p ≥ 0.
p
Since p ≤ 3(k − 1)2 /(k2 − 2k) and k ≥ 2 + 3, we have
PI
M
9(k − 1)4 9(k − 1)2
3(k − 1)2 − p2 − 3p ≥ 3(k − 1)2 − −
LY
k2 (k − 2)2 k(k − 2)
3(k − 1) (k − 3)(k − 4k + 1)
2 2 2
=
O
≥ 0.
k2 (k − 2)2
p
.M
a = b = c = 1. In the case k = 2 + 3, the equality holds again

The equality holds for p
for a = 0 and b = c = 3 (or any cyclic permutation).
W
W
P 1.39. Let a, b, c be nonnegative real numbers such that a2 + b2 + c 2 = 3. Prove that

W
a(b + c) b(c + a) c(a + b)

+ + ≤ 3.
1 + bc 1 + ca 1 + ab
Solution. Write the inequality in the homogeneous form

X a(b + c)
≤ 1,
a2 + b2 + c 2 + 3bc
or X a(b + c)

a
− ≤ 0,
a2 + b2 + c 2 + 3bc a + b + c
74 Vasile Cîrtoaje
X a(a − b)(a − c)
≥ 0.
a2 + b2 + c 2 + 3bc
Without loss of generality, assume that a ≥ b ≥ c. Then, it suffices to prove that
a(a − b)(a − c) b(b − c)(b − a)
+ 2 ≥ 0.
a2 + b + c + 3bc a + b2 + c 2 + 3ca
2 2
This is true if
a(a − c) b(b − c)
≥ 2 .
a2+ b + c + 3bc
2 2 a + b2 + c 2 + 3ca
Since a(a − c) ≥ b(b − c) and
1 1
L
≥ 2 ,
a2 + b2 + c + 3bc
2 a + b + c 2 + 3ca
2
.M
the conclusion
p follows. The equality holds for a = b = c = 1, and for a = 0 and
b = c = 3/2 (or any cyclic permutation).
D
A
PI
P 1.40. Let a, b, c be positive real numbers such that a2 + b2 + c 2 = 3. Prove that
a2 + b2 b2 + c 2 c 2 + a2
M
+ + ≤ 3.
a+b b+c c+a
LY
(Cezar Lupu, 2005)

O
First Solution. We apply the SOS method. Write the inequality in the homogeneous
form
X b2 + c 2 b + c Æ
.M
− ≥ 3(a2 + b2 + c 2 ) − a − b − c,
b+c 2
or
W
(b − c)2
P
X (b − c)2
≥p .
2(b + c) 3(a2 + b2 + c 2 ) + a + b + c
W
p
Since 3(a2 + b2 + c 2 ) + a + b + c ≥ 2(a + b + c) > 2(b + c), the conclusion follows.
W

Second Solution. By virtue of the Cauchy-Schwarz inequality, we have
Pp Pp
( a2 + b2 )2 2 a2 + 2 (a2 + b2 )(a2 + c 2 )
P
X a2 + b2
≥ =
a+b (a + b)
P P
2 a
2 a + 2 (a + bc) 3 a + ( a)2
P 2 P 2 P 2 P
≥ P = P
2 a 2 a
9 + ( a)2 ( a − 3)2
P P
= P =3+ P ≥ 3.
2 a 2 a
ab bc ca 7
+ + + 2 ≤ (a + b + c).
a+b b+c c+a 6
Solution. We apply the SOS method. Write the inequality as

X 4bc

3 b+c− ≥ 8(3 − a − b − c).
b+c
Since
L
4bc (b − c)2
b+c− =
.M
b+c b+c
and
D
9 − (a + b + c)2 3(a2 + b2 + c 2 ) − (a + b + c)2
3−a− b−c = =
A
3+a+ b+c 3+a+ b+c
=
1
3+a+ b+c
PI X
(b − c)2 ,
M
LY
Sa (b − c)2 + S b (c − a)2 + Sc (a − b)2 ≥ 0,

O
where
3 8
Sa = − .
.M
b+c 3+a+ b+c

Without loss of generality, assume that a ≥ b ≥ c. Since Sa ≥ S b ≥ Sc , it suffices to show
that S b + Sc ≥ 0. Indeed, if this condition is fulfilled, then Sa ≥ S b ≥ (S b + Sc )/2 ≥ 0,
W
and hence
W
Sa (b − c)2 + S b (c − a)2 + Sc (a − b)2 ≥ S b (c − a)2 + Sc (a − b)2

W
≥ S b (a − b)2 + Sc (a − b)2 = (a − b)2 (S b + Sc ) ≥ 0.

Since
4(9 − 5a − b − c)
S b + Sc = ,
(a + b + 2c)(3 + a + b + c)
we need to show that 9 ≥ 5a+ b+c. This follows immediately from the Cauchy-Schwarz
inequality
(25 + 1 + 1)(a2 + b2 + c 2 ) ≥ (5a + b + c)2 .
Thus, the proof is completed. The equality holds for a = b = c = 1, and for a = 5/3
and b = c = 1/3 (or any cyclic permutation).
76 Vasile Cîrtoaje
1 1 1 3
(a) + + ≤ ;
3 − a b 3 − bc 3 − ca 2
1 1 1 3
(b) p +p +p ≤p .
6 − ab 6 − bc 6 − ca 6−1
Solution. (a) Since
L
3 ab 2a b
=1+ =1+ 2
.M
3 − ab 3 − ab a + b + 2c 2 + (a − b)2
2
2a b (a + b)2
≤1+ 2 ≤ 1 + ,
D
a + b2 + 2c 2 2(a2 + b2 + 2c 2 )
A
(a + b)2 (b + c)2
PI
(c + a)2
+ + ≤ 3.
M
a2 + b2 + 2c 2 b2 + c 2 + 2a2 c 2 + a2 + 2b2
LY
(a + b)2 (a + b)2 a2 b2
O
= ≤ + .
a2 + b2 + 2c 2 (a2 + c 2 ) + (b2 + c 2 ) a2 + c 2 b2 + c 2
.M
Thus,
(a + b)2 X a2 X b2 X a2 X c2
W
X
≤ + = + = 3.
a2 + b2 + 2c 2 a2 + c 2 b2 + c 2 a2 + c 2 c 2 + a2
W

W
(b) According to P 1.29, the following inequality holds
1 1 1 3
2 2
+ 2 2
+ 2 2
≤ .
6−a b 6− b c 6−c a 5
Since p
2 6 1 1
2 2
=p +p ,
6−a b 6 − ab 6 + ab
this inequality becomes
p
X 1 X 1 6 6
p + p ≤ .
6 − ab 6 + ab 5

X 1 3
p ≥p .
6 + ab 6+1
Since a b + bc + ca ≤ a2 + b2 + c 2 = 3, by the Cauchy-Schwarz inequality, we have
X 1 9 9 3
p ≥Pp = p ≥p .
6 + ab ( 6 + a b) 3 6 + a b + bc + ca 6+1
L
.M
D
1 1 1 3
+ + ≥ .
1+a 5 1+ b 5 1+c 5
A
2
Solution. Let a = min{a, b, c}. There are two cases to consider

M
Case 1: a ≥ 1/2. The desired inequality follows by summing the inequalities
LY
8 8 8
≥ 9 − 5a2 , ≥ 9 − 5b2 , ≥ 9 − 5c 2 ;
1 + a5 1 + b5 1 + c5
O
8
.M
To obtain these inequalities, we start from the inequality ≥ p + q x 2 , whose

1 + x5
coefficients p and q will be determined such that the polynomial P(x) = 8−(1+ x 5 )(p +
q x 2 ) divides by (x − 1)2 . It is easy to check that P(1) = 0 involves p + q = 4, when
W
P(x) = 4(2 − x 2 − x 7 ) − p(1 − x 2 + x 5 − x 7 ) = (1 − x)Q(x),

W
where
W
Q(x) = 4(2 + 2x + x 2 + x 3 + x 4 + x 5 + x 6 ) − p(1 + x + x 5 + x 6 ).

In addition, Q(1) = 0 involves p = 9. In this case,
P(x) = (1 − x)2 (5x 5 + 10x 4 + 6x 3 + 2x 2 − 2x − 1)

= (1 − x)2 [x 5 + (2x − 1)(2x 4 + 6x 3 + 6x 2 + 4x + 1)] ≥ 0.
Case 2: a ≥ 1/2. Write the inequality as
1 1 b5 c 5 − 1
− ≥ .
1 + a5 2 (1 + b5 )(1 + c 5 )
78 Vasile Cîrtoaje
Since
1 1 32 1 31
− ≥ − =
1+a 5 2 33 2 66
and p
(1 + b5 )(1 + c 5 ) ≥ (1 + b5 c 5 )2 ,
it suffices to show that p
31(1 + b5 c 5 )2 ≥ 66(b5 c 5 − 1),
which is equivalent to bc ≤ (97/35)2/5 . Indeed, from
3 = a2 + b2 + c 2 > b2 + c 2 ≥ 2bc,
L
we get bc < 3/2 < (97/35)2/5 . This completes the proof. The equality holds for a =
.M
b = c = 1.
D
A
PI
P 1.44. Let a, b, c be positive real numbers such that a bc = 1. Prove that
1 1 1
M
+ + ≥ 1.
a2 + a + 1 b2 + b + 1 c 2 + c + 1
LY
First Solution. Using the substitutions a = yz/x 2 , b = z x/ y 2 , c = x y/z 2 , where x, y, z

O
are positive real numbers, the inequality becomes

.M
X x4
≥ 1.
x 4 + x 2 yz + y 2 z 2
W

W
( x 2 )2 x + 2 y 2z2
P P 4 P
X x4
≥P =P .
W
x 4 + x 2 yz + y 2 z 2 (x 4 + x 2 yz + y 2 z 2 ) x 4 + x yz x + y 2 z 2
P P

X X
y 2 z 2 ≥ x yz x,
x 2 ( y − z)2 ≥ 0. The equality holds for a = b = c = 1.

P
Second Solution. Using the substitutions a = y/x, b = z/ y, c = x/z, where x, y, z > 0,
x2 y2 z2
+ + ≥ 1.
x2 + x y + y2 y 2 + yz + z 2 z 2 + z x + z 2
Since
x 2 (x 2 + y 2 + z 2 + x y + yz + z x) 2 x 2 z(x + y + z)
= x + ,
x2 + x y + y2 x2 + x y + y2
multiplying by x 2 + y 2 + z 2 + x y + yz + z x, the inequality can be written as
X x 2z x y + yz + z x
≥ .
x2 + x y + y2 x + y +z
( xz)2
P
X x 2z x y + yz + z x
≥P = .
L
x +xy+ y
2 2 z(x + x y + y )
2 2 x + y +z
.M
Remark. The inequality in P 1.44 is a particular case of the following more general
inequality: (Vasile Cîrtoaje, 2009).
D
• Let a1 , a2 , . . . , an (n ≥ 3) be positive real numbers such that a1 a2 · · · an = 1. If p and
A
q are nonnegative real numbers satisfying p + q = n − 1, then
i=n
X 1
PI
≥ 1.
M
i=1
1 + pai + qai2
LY
O

.M
1 1 1
+ 2 + 2 ≤ 3.
a2 −a+1 b − b+1 c −c+1
W
First Solution. Since

W
1 1 2(a2 + 1) 2a4
W
+ = = 2 − ,
a2 − a + 1 a2 + a + 1 a4 + a2 + 1 a4 + a2 + 1
X 1 X a4
+ 2 ≥ 3.
a2 + a + 1 a4 + a2 + 1
X 1
≥1
a2 + a + 1
and
X a4
≥ 1.
a4 + a2 + 1
80 Vasile Cîrtoaje
The first inequality is just the inequality in P 1.44, while the second follows from the
first by substituting a, b, c with a−2 , b−2 , c −2 , respectively. The equality holds for a =
b = c = 1.
Second Solution. Write the inequality as
X4 1

− 2 ≥ 1,
3 a −a+1
X (2a − 1)2
≥ 3.
a2 − a + 1
Let p = a + b + c and q = a b + bc + ca. By the Cauchy-Schwarz inequality, we have
L
(2 a − 3)2
P
X (2a − 1)2 (2p − 3)2
≥ = .
.M
a2 − a + 1 (a2 − a + 1) p2 − 2q − p + 3
P
D
(2p − 3)2 ≥ 3(p2 − 2q − p + 3),
A
which is equivalent to PI
p2 + 6q − 9p ≥ 0.
M
From the known inequality
LY
(a b + bc + ca)2 ≥ 3a bc(a + b + c),

we get q2 ≥ 3p. Using this inequality and the AM-GM inequality, we get
O
Æ3
Æ
3
p2 + 6q = p2 + 3q + 3q ≥ 3 9p2 q2 ≥ 3 9p2 (3p) = 9p.
.M
W

W
3+a 3+ b 3+c
+ + ≥ 3.
(1 + a)2 (1 + b)2 (1 + c)2
W
Solution. Using the inequality in P 1.1, we have

X 3+a X 2 X 1
= +
(1 + a)2 (1 + a)2 1+a
X 1 1 1
X
= + +
(1 + a) 2 (1 + b)2 1+c
X 1 X ab
≥ + = 3.
1 + ab 1 + ab

7 − 6a 7 − 6b 7 − 6c
+ + ≥ 1.
2 + a2 2 + b2 2 + c 2
Solution. Write the inequality as
7 − 6a 7 − 6b 7 − 6c

+1 + +1 + + 1 ≥ 4,
2 + a2 2 + b2 2 + c2
(3 − a)2 (3 − b)2 (3 − c)2
+ + ≥ 4.
2 + a2 2 + b2 2 + c2
L
Substituting a, b, c by 1/a, 1/b, 1/c, respectively, we need to prove that a bc = 1 involves
.M
(3a − 1)2 (3b − 1)2 (3c − 1)2
+ + ≥ 4.
2a2 + 1 2b2 + 1 2c 2 + 1
D
A
(3 a − 3)2 9 a2 + 18 a b − 18 a + 9
P P P P
X (3a − 1)2
2a2 + 1
≥ P
(2a2 + 1)
= PI
2 a2 + 3
P .
M
Thus, it suffices to prove that
LY
f (a) + f (b) + f (c) ≥ 3,

where
O
1

2
f (x) = x + 18 −x .
x
.M
In order to do this, we use the mixing method. Without loss of generality, assume that
a = max{a, b, c}, a ≥ 1, bc ≤ 1. Since
W
p 1
p p
f (b) + f (c) − 2 f ( bc) = (b − c)2 + 18( b − c)2 − 1 ≥ 0,
bc
W
it suffices to show that p

f (a) + 2 f ( bc) ≥ 3.
W

1

2
f (x ) + 2 f ≥ 3,
x
p
where x = a. It is equivalent to
x 6 − 18x 4 + 36x 3 − 3x 2 − 36x + 20 ≥ 0,
(x − 1)2 (x − 2)2 (x + 1)(x + 5) ≥ 0.
Since the last inequality is true, the proof is completed. The equality holds for a = b =
c = 1, and also for a = 1/4 and b = c = 2 (or any cyclic permutation).
82 Vasile Cîrtoaje
a6 b6 c6
+ + ≥ 1.
1 + 2a5 1 + 2b5 1 + 2c 5
Solution. Using the substitutions

v v v
2 3 y
2 2
3 x 3 z
t t t
a= , b= , c= ,
yz zx xy
L
the inequality becomes
.M
X x4
p ≥ 1.
y 2 z 2 + 2x 3 3 x yz
D
A
( x 2 )2 ( x 2 )2
P P
x4
X
p
y 2 z 2 + 2x 3 3 x yz
≥P p
( y 2 z 2 + 2x 3 3 x yz)
=PPI p P .
x 2 y 2 + 2 3 x yz x 3
M
Therefore, we need to show that
LY
X X p X
( x 2 )2 ≥ x 2 y 2 + 2 3 x yz x 3.
O
p
Since x + y + z ≥ 3 3 x yz, it suffices to prove that
.M
X X X X
3( x 2 )2 ≥ 3 x 2 y 2 + 2( x)( x 3 );
W
that is, X X X
x4 + 3 x2 y2 ≥ 2 x y(x 2 + y 2 ),
W
or, equivalently, X
W
(x − y)4 ≥ 0.
a b c 1
+ + ≤ .
a2 + 5 b2 + 5 c 2 + 5 2
Solution. Let
a b c
F (a, b, c) = + 2 + 2 .
a2 +5 b +5 c +5
Without loss of generality, assume that a = min{a, b, c}.
Case 1: a ≤ 1/5. We have
a b c 1 1 1
F (a, b, c) <
+ p + p ≤ +p < .
5 2 5b2 2 5c 2 25 5 2
p p
Case 2: a > 1/5. Let x = bc, a = 1/x 2 , x < 5. We will show that
L
1
.M
F (a, b, c) ≤ F (a, x, x) ≤ .
2
The left inequality, F (a, b, c) ≤ F (a, x, x), is equivalent to
D
p
A
p
( b − c)2 [10x(b + c) + 10x 2 − 25 − x 4 ] ≥ 0.
This is true since

PI
M
10x(b + c) + 10x 2 − 25 − x 4 ≥ 20x 2 + 10x 2 − 25x 2 − x 4 = x 2 (5 − x 2 ) > 0.
LY
The right inequality, F (a, x, x) ≤ 1/2, is equivalent to

O
(x − 1)2 (5x 4 − 10x 3 − 2x 2 + 6x + 5) ≥ 0.

.M
It is also true since

W
5x 4 − 10x 3 − 2x 2 + 6x + 5 = 5(x − 1)4 + 2x(5x 2 − 16x + 13)

W
and p
5x 2 + 13 ≥ 2 65x 2 > 16x.
W
1 1 1 2
+ + + ≥ 1.
(1 + a)2 (1 + b)2 (1 + c)2 (1 + a)(1 + b)(1 + c)
(Pham Van Thuan, 2006)

84 Vasile Cîrtoaje
First Solution. There are two of a, b, c either greater than or equal to 1, or less than or
equal to 1. Let b and c be these numbers; that is, (1 − b)(1 − c) ≥ 0. Since
1 1 1
+ ≥
(1 + b)2 (1 + c)2 1 + bc
(see P 1.1), it suffices to show that
1 1 2
+ + ≥ 1.
(1 + a)2 1 + bc (1 + a)(1 + b)(1 + c)
L
b2 c 2 1 2bc
+ +
.M
≥ 1,
(1 + bc) 2 1 + bc (1 + bc)(1 + b)(1 + c)
which can be written in the obvious form
D
bc(1 − b)(1 − c)
A
≥ 0.
(1 + bc)(1 + b)(1 + c)
PI
M
Second Solution. Setting a = yz/x 2 , b = z x/ y 2 , c = x y/z 2 , where x, y, z > 0, the
LY
inequality becomes
X x4 2x 2 y 2 z 2
O
+ ≥ 1.
(x 2 + yz)2 (x 2 + yz)( y 2 + z x)(z 2 + x y)
.M
Since (x 2 + yz)2 ≤ (x 2 + y 2 )(x 2 + z 2 ), we have

X x4 X x4 2x 2 y 2 z 2
W
≥ = 1 − .
(x 2 + yz)2 (x 2 + y 2 )(x 2 + z 2 ) (x 2 + y 2 )( y 2 + z 2 )(z 2 + x 2 )
W

W
(x 2 + y 2 )( y 2 + z 2 )(z 2 + x 2 ) ≥ (x 2 + yz)( y 2 + z x)(z 2 + x y).
This inequality follows by multiplying the inequalities
(x 2 + y 2 )(x 2 + z 2 ) ≥ (x 2 + yz)2 ,
( y 2 + z 2 )( y 2 + x 2 ) ≥ ( y 2 + z x)2 ,
(z 2 + x 2 )(z 2 + y 2 ) ≥ (z 2 + x y)2 .
Third Solution. We make the substitutions
1 1+ x 1 1+ y 1 1+z
= , = , = ;
1+a 2 1+ b 2 1+c 2
that is,
1− x 1− y 1−z
a= , b= , c= ,
1+ x 1+ y 1+z
where −1 < x, y, z < 1. Since a bc = 1 involves x + y + z + x yz = 0, we need to prove
that
x + y + z + x yz = 0
implies
(1 + x)2 + (1 + y)2 + (1 + z)2 + (1 + x)(1 + y)(1 + z) ≥ 4.
L
x 2 + y 2 + z 2 + (x + y + z)2 + 4(x + y + z) ≥ 0.
.M
By virtue of the AM-GM inequality, we have
D
x 2 + y 2 + z 2 + (x + y + z)2 + 4(x + y + z) = x 2 + y 2 + z 2 + x 2 y 2 z 2 − 4x yz
A
Æ4
≥ 4 x 4 y 4 z 4 − 4x yz = 4|x yz| − 4x yz ≥ 0.
PI
M
LY
P 1.51. Let a, b, c be nonnegative real numbers such that

O
1 1 1 3
+ + = .
a+b b+c c+a 2
.M
Prove that
3 2 1
≥ + 2 .
W
a+b+c a b + bc + ca a + b2 + c 2
W

W
2 1 1 1 2 1

+ + ≥ + 2 .
a+b+c a+b b+c c+a a b + bc + ca a + b2 + c 2
Denote q = a b + bc + ca and assume that a + b + c = 1. From the known inequality
(a + b + c)2 ≥ 3(a b + bc + ca), we get 1 − 3q ≥ 0. Rewrite the desired inequality as
follows
1 1 1 2 1

2 + + ≥ + ,
1−c 1−a 1− b q 1 − 2q
2(q + 1) 2 − 3q
≥ ,
q − a bc q(1 − 2q)
q2 (1 − 4q) + (2 − 3q)a bc ≥ 0.
86 Vasile Cîrtoaje
By Schur’s inequality, we have
1 − 4q ≥ −9a bc.
Then,
q2 (1 − 4q) + (2 − 3q)a bc ≥ −9q2 a bc + (2 − 3q)a bc

= (1 − 3q)(2 + 3q)a bc ≥ 0.
The equality holds for a = b = c = 1, and for a = 0 and b = c = 5/3 (or any cyclic
L
permutation).
D
.M
A
P 1.52. Let a, b, c be nonnegative real numbers such that
PI
7(a2 + b2 + c 2 ) = 11(a b + bc + ca).
M
Prove that
51 a b c
+ +
LY
≤ ≤ 2.
28 b+c c+a a+b
O
Solution. Due to homogeneity, we may assume that b + c = 2. Let us denote x = bc,

0 ≤ x ≤ 1. By the hypothesis 7(a2 + b2 + c 2 ) = 11(a b + bc + ca), we get
.M
7a2 − 22a + 28
x= .
W
25
Then, x ≤ 1 involves 1/7 ≤ a ≤ 3. Since
W
a b c a a(b + c) + (b + c)2 − 2bc

W
+ + = +
b+c c+a a+b b+c a2 + (b + c)a + bc
a 2(a + 2 − x) 4a3 + 27a + 11
= + 2 = ,
2 a + 2a + x 8a2 + 7a + 7
the required inequalities become
51 4a3 + 27a + 11
≤ ≤ 2.
28 8a2 + 7a + 7
We have
4a3 + 27a + 11 51 (7a − 1)(4a − 7)2
− = ≥0
8a2 + 7a + 7 28 28(8a2 + 7a + 7)
and
4a3 + 27a + 11 (3 − a)(2a − 1)2
2− = ≥ 0.
8a2 + 7a + 7 8a2 + 7a + 7
This completes the proof. The left inequality becomes an equality for 7a = b = c (or
any cyclic permutation), while the right inequality is an equality for a/3 = b = c (or
any cyclic permutation).
L
1 1 1 10
+ 2 + 2 ≥ .
.M
a2 +b 2 b +c 2 c +a 2 (a + b + c)2
D
Solution. Assume that a = min{a, b, c}, and denote
A
a a
x = b+ , y =c+ .
2 2 PI
Since
M
a2 + b2 ≤ x 2 , b2 + c 2 ≤ x 2 + y 2 , c 2 + a2 ≤ y 2 ,
LY
(a + b + c)2 = (x + y)2 ≥ 4x y,
O

1 1 1 5
+ 2 + 2≥ .
x +y
.M
x 2 2 y 2x y
We have
W
1 1 1 5 1 1 2 1 1

+ 2 + 2− = + 2− + −
x 2 x +y 2 y 2x y x2 y xy x 2 + y 2 2x y
W
(x − y)2 (x − y)2
= −
x2 y2 2x y(x 2 + y 2 )
W
(x − y) (2x 2 − x y + 2 y 2 )
2
= ≥ 0.
2x 2 y 2 (x 2 + y 2 )
The equality holds for a = 0 and b = c (or any cyclic permutation).
1 1 1 3
+ 2 + 2 ≥ .
a2 − ab + b 2 b − bc + c 2 c − ca + a 2 max{a b, bc, ca}
88 Vasile Cîrtoaje
Solution. Assume that a = min{a, b, c}, hence bc = max{a b, bc, ca}. Since
1 1 1 1 1 1
+ 2 + 2 ≥ 2+ 2 + 2,
a2 − ab + b 2 b − bc + c 2 c − ca + a 2 b b − bc + c 2 c
1 1 1 3
+ + ≥ .
b2 b2 − bc + c 2 c 2 bc
We have
1 1 1 3 (b − c)4
+ + − = ≥ 0.
b2 b2 − bc + c 2 c 2 bc b2 c 2 (b2 − bc + c 2 )
The equality holds for a = b = c, and also a = 0 and b = c (or any cyclic permutation).
L
.M
D
A
a(2a + b + c) b(2b + c + a) c(2c + a + b)
+ + PI ≥ 6.
b2 + c 2 c 2 + a2 a2 + b2
M
Solution. By the Cauchy-Schwarz inequality, we have
LY
[ a(2a + b + c)]2
P
X a(2a + b + c)
≥P .
O
b2 + c 2 a(2a + b + c)(b2 + c 2 )
.M
Thus, we still need to show that

X X X
2( a2 + a b)2 ≥ 3 a(2a + b + c)(b2 + c 2 ),
W
W
X X X X
2 a4 + 2a bc a+ a b(a2 + b2 ) ≥ 6 a2 b2 .
W
We can obtain this inequality by adding Schur’s inequality of degree four

X X X
a4 + a bc a≥ a b(a2 + b2 )
and X X
a b(a2 + b2 ) ≥ 2 a2 b2 ,
multiplied by 2 and 3, respectively. The equality occurs for a = b = c, and for a = 0
and b = c (or any cyclic permutation).
a2 (b + c)2 b2 (c + a)2 c 2 (a + b)2

+ 2 + ≥ 2(a b + bc + ca).
b2 + c 2 c + a2 a2 + b2
Solution. We apply the SOS method. Since
a2 (b + c)2 ) 2 2a2 bc
= a + ,
b2 + c 2 b2 + c 2
L
.M
X 2bc
X X
2( a2 − a b) − a2 1 − ≥ 0,
b2 + c 2
D
X X a2 (b − c)2
(b − c)2 − ≥ 0,
A
b2 + c 2
X
1− 2
a2

PI
(b − c)2 ≥ 0.
b + c2
M
c2
> 0, it suffices to
LY
Without loss of generality, assume that a ≥ b ≥ c. Since 1 −

a2 + b2
prove that
O
a2 b2

2
1− 2 (b − c) + 1 − 2 (a − c)2 ≥ 0,
b + c2 c + a2
.M
(a2 − b2 + c 2 )(a − c)2 (a2 − b2 − c 2 )(b − c)2

W
≥ .
a2 + c 2 b2 + c 2
W
This inequality follows by multiplying the inequalities

W
(a − c)2 (b − c)2
a2 − b2 + c 2 ≥ a2 − b2 − c 2 , ≥ .
a2 + c 2 b2 + c 2
The latter inequality is true since
(a − c)2 (b − c)2 2bc 2ac 2c(a − b)(a b − c 2 )

− = − = ≥ 0.
a2 + c 2 b2 + c 2 b2 + c 2 a2 + c 2 (b2 + c 2 )(a2 + c 2 )
The equality occurs for a = b = c, and for a = 0 and b = c (or any cyclic permutation).
90 Vasile Cîrtoaje
P 1.57. If a, b, c are real numbers such that a bc > 0, then

1 1 1

X a a b c
3 + 5 + + ≥ 8 + + .
b2 − bc + c 2 bc ca a b a b c
Solution. In order to apply the SOS method, we multiply the inequality by a bc and
write it as follows:
X X X bc

2 2
8( a − bc) − 3 a 1− 2 ≥ 0,
b − bc + c 2
X a2 (b − c)2
L
X
4 (b − c)2 − 3 ≥ 0,
b2 − bc + c 2
.M
X (b − c)2 (4b2 − 4bc + 4c 2 − 3a2 )
≥ 0.
b2 − bc + c 2
D
Without loss of generality, assume that a ≥ b ≥ c. Since
A
4a2 − 4a b + 4b2 − 3c 2 = (2a − b)2 + 3(b2 − c 2 ) ≥ 0,
PI
M
(a − c)2 (4a2 − 4ac + 4c 2 − 3b2 ) (b − c)2 (3a2 − 4b2 + 4bc − 4c 2 )
LY
≥ .
a2 − ac + c 2 b2 − bc + c 2
Notice that
O
4a2 − 4ac + 4c 2 − 3b2 = (a − 2c)2 + 3(a2 − b2 ) ≥ 0.

.M
Thus, the desired inequality follows by multiplying the inequalities
4a2 − 4ac + 4c 2 − 3b2 ≥ 3a2 − 4b2 + 4bc − 4c 2

W
and
(a − c)2 (b − c)2
W
≥ .
a2 − ac + c 2 b2 − bc + c 2
W
The first inequality is equivalent to
(a − 2c)2 + (b − 2c)2 ≥ 0.
Also, we have
(a − c)2 (b − c)2 bc ac
− = 2 − 2
a − ac + c
2 2 b − bc + c
2 2 b − bc + c 2 a − ac + c 2
c(a − b)(a b − c 2 )
= 2 ≥ 0.
(b − bc + c 2 )(a2 − ac + c 2 )
The equality occurs for a = b = c, and for 2a = b = c (or any cyclic permutation).
1 1 1

(a) 2a bc + + + a2 + b2 + c 2 ≥ 2(a b + bc + ca);
a+b b+c c+a
a2 b2 c2 3(a2 + b2 + c 2 )
(b) + + ≤ .
a+b b+c c+a 2(a + b + c)
Solution. (a) First Solution. We have

X 1 X X a(2bc + a b + ac)
2a bc + a2 =
b+c b+c
L
X a b(a + c) X ac(a + b)
.M
= +
b+c b+c
X a b(a + c) X ba(b + c)
= +
D
b+ c c + a X
a+c b+c
A
X
= ab + ≥2 a b.
b+c a+c PI
M
LY
X 2a bc
2
+ a − a b − ac ≥ 0.
b+c
O
We have
.M
X 2a bc X a b(a − b) + ac(a − c)
2
+ a − a b − ac =
b+c b+c
W
= +
b+c c+a
W
X a b(a − b)2
= ≥ 0.
(b + c)(c + a)
W
(b) Since
X a2 X ab
X ab
= a− =a+b+c− ,
a+b a+b a+b
we can write the desired inequality as
X ab 3(a2 + b2 + c 2 )
+ ≥ a + b + c.
a+b 2(a + b + c)
Multiplying by 2(a + b + c), the inequality can be written as
X a
2 1+ bc + 3(a2 + b2 + c 2 ) ≥ 2(a + b + c)2 ,
b+c
92 Vasile Cîrtoaje
or X 1
2a bc + a2 + b2 + c 2 ≥ 2(a b + bc + ca),
b+c
which is just the inequality in (a).
a2 − bc b2 − ca c 2 − a b 3(a b + bc + ca)
(a) + 2 + + ≥ 3;
b2 + c 2 c + a2 a2 + b2 a2 + b2 + c 2
L
a2 b2 c2 a b + bc + ca 5
.M
(b) + + + 2 ≥ ;
b +c
2 2 c +a
2 2 a +b
2 2 a +b +c2 2 2
D
a2 + bc b2 + ca c 2 + a b a b + bc + ca
(c) + 2 + 2 ≥ 2 + 2.
b +c
2 2 c +a 2 a +b 2 a + b2 + c 2
A
Solution. (a) Write the inequality as follows:

M
X 2a2 a b + bc + ca
X
2bc

LY
− 1 + 1 − − 6 1 − ≥ 0,
b2 + c 2 b2 + c 2 a2 + b2 + c 2
O
X 2a2 − b2 − c 2 X (b − c)2 X (b − c)2

+ −3 ≥ 0.
b2 + c 2 b2 + c 2 a2 + b2 + c 2
.M
Since
X 2a2 − b2 − c 2 X a2 − b2 X a2 − c 2 X a2 − b2 X b2 − a2
W
= + = +
b2 + c 2 b2 + c 2 b2 + c 2 b2 + c 2 c 2 + a2
(a − b )
2 2 2
(b2 − c 2 )2
W
X X
= = ,
(b2 + c 2 )(c 2 + a2 ) (a2 + b2 )(a2 + c 2 )
W

X
(b − c)2 Sa ≥ 0,
where
(b + c)2 1 3
Sa = + 2 − 2 .
(a + b )(a + c ) b + c
2 2 2 2 2 a + b2 + c 2
It suffices to show that Sa ≥ 0 for all nonnegative real numbers a, b, c, no two of which
are zero. Denoting x 2 = b2 + c 2 , we have
x 2 + 2bc 1 3
Sa = + 2− 2 ,
a +a x +b c
4 2 2 2 2 x a + x2
and the inequality Sa ≥ 0 becomes
(a2 − 2x 2 )b2 c 2 + 2x 2 (a2 + x 2 )bc + (a2 + x 2 )(a2 − x 2 )2 ≥ 0.
Clearly, this is true if

−2x 2 b2 c 2 + 2x 4 bc ≥ 0.
Indeed,
−2x 2 b2 c 2 + 2x 4 bc = 2x 2 bc(x 2 − bc) = 2bc(b2 + c 2 )(b2 + c 2 − bc) ≥ 0.
L
.M
(b) First Solution. We get the desired inequality by summing the inequality in (a)
and the inequality
D
bc ca ab 1 2(a b + bc + ca)
+ 2 + 2 + ≥ .
+c c +a a +b a2 + b2 + c 2
A
b2 2 2 2 2
This inequality is equivalent to PI
X 2bc 4(a b + bc + ca)

M
+ 1 ≥ + 2,
b2 + c 2 a2 + b2 + c 2
LY
X (b + c)2 2(a + b + c)2

≥ .
b2 + c 2 a2 + b2 + c 2
O

.M
P 2
X (b + c)2 (b + c) 2(a + b + c)2
≥ P = 2 .
b2 + c 2 (b2 + c 2 ) a + b2 + c 2
W
W
Second Solution. Let

W

P 2 2
X a2 a (p2 − 2q)2
≥ = .
b2 + c 2 a2 (b2 + c 2 ) 2(q2 − 2pr)
P
(p2 − 2q)2 2q
2
+ 2 ≥ 5. (*)
q − 2pr p − 2q
94 Vasile Cîrtoaje
Consider the following cases: p2 ≥ 4q and 3q ≤ p2 < 4q.

Case 1: p2 ≥ 4q. The inequality (*) is true if
(p2 − 2q)2 2q
2
+ 2 ≥ 5,
q p − 2q
(p2 − 4q) (p2 − q)2 − 2q2 ≥ 0.

Case 2: 3q ≤ p2 < 4q. Using Schur’s inequality of degree four
L
.M
6pr ≥ (p2 − q)(4q − p2 ),
the inequality (*) is true if
D
A
3(p2 − 2q)2 2q
+ ≥ 5,
3q2 − (p2 − q)(4q − p2 ) p2 − 2qPI
M
(p2 − 3q)(p2 − 4q)(2p2 − 5q) ≤ 0.
LY
Third Solution (by Nguyen Van Quy). Write the inequality (*) from the preceding solu-
O
tion as follows:
(a2 + b2 + c 2 )2 2(a b + bc + ca)
.M
+ ≥ 5,
a b +b c +c a
2 2 2 2 2 2 a2 + b2 + c 2
(a2 + b2 + c 2 )2 2(a b + bc + ca)
W
−3≥2− ,
a b +b c +c a
2 2 2 2 2 2 a2 + b2 + c 2
W
a4 + b4 + c 4 − a2 b2 − b2 c 2 − c 2 a2 2(a2 + b2 + c 2 − a b − bc − ca)
≥ .
a2 b2 + b2 c 2 + c 2 a2 a2 + b2 + c 2
W
Since
X
2(a2 b2 + b2 c 2 + c 2 a2 ) ≤ a b(a2 + b2 ) ≤ (a b + bc + ca)(a2 + b2 + c 2 ),
a4 + b4 + c 4 − a2 b2 − b2 c 2 − c 2 a2
≥ a2 + b2 + c 2 − a b − bc − ca,
a b + bc + ca
which is just Schur’s inequality of degree four
a4 + b4 + c 4 + a bc(a + b + c) ≥ a b(a2 + b2 ) + bc(b2 + c 2 ) + ca(c 2 + a2 ).

(c) We get the desired inequality by summing the inequality in (a) and the inequal-
ity
2bc 2ca 2a b 4(a b + bc + ca)
+ 2 + 2 +1≥ ,
+cb22 c +a 2 a +b 2 a2 + b2 + c 2
which is proved at (b). The equality occurs for a = b = c, and for a = 0 and b = c (or
L
a2 b2 c2 (a + b + c)2
+ +
.M
≥ .
b2 + c 2 c 2 + a2 a2 + b2 2(a b + bc + ca)
D
Solution. Applying the Cauchy-Schwarz inequality, we have
A
P 2 2
X a2 a (a2 + b2 + c 2 )2
PI
≥ = .
b2 + c 2 a2 (b2 + c 2 ) 2(a2 b2 + b2 c 2 + c 2 a2 )
P
M
LY
(a2 + b2 + c 2 )2 (a + b + c)2
≥ ,
2(a2 b2 + b2 c 2 + c 2 a2 ) 2(a b + bc + ca)
O
.M
(a2 + b2 + c 2 )2 (a + b + c)2
− 3 ≥ − 3,
a2 b2 + b2 c 2 + c 2 a2 a b + bc + ca
W
a4 + b4 + c 4 − a2 b2 − b2 c 2 − c 2 a2 a2 + b2 + c 2 − a b − bc − ca
W
≥ .
a2 b2 + b2 c 2 + c 2 a2 a b + bc + ca
W
Since a2 b2 + b2 c 2 + c 2 a2 ≤ (a b + bc + ca)2 , it suffices to show that
a4 + b4 + c 4 − a2 b2 − b2 c 2 − c 2 a2 ≥ (a2 + b2 + c 2 − a b − bc − ca)(a b + bc + ca),
which is just Schur’s inequality of degree four
a4 + b4 + c 4 + a bc(a + b + c) ≥ a b(a2 + b2 ) + bc(b2 + c 2 ) + ca(c 2 + a2 ).
tion).
96 Vasile Cîrtoaje
2a b 2bc 2ca a2 + b2 + c 2 5
+ + + ≥ .
(a + b)2 (b + c)2 (c + a)2 a b + bc + ca 2
First Solution. We use the SOS method. Write the inequality as follows
a2 + b2 + c 2 X1 2bc

−1≥ − ,
a b + bc + ca 2 (b + c)2
L
X (b − c)2 X (b − c)2
≥ ,
.M
ab + bc + ca (b + c)2
(b − c)2 Sa + (c − a)2 S b + (a − b)2 Sc ≥ 0,
D
where
A
Sa = 1 −
(b + c)2
, Sb = 1 −
(c + a)2
PI
, Sc = 1 −
(a + b)2
.
M
Without loss of generality, assume that a ≥ b ≥ c. We have Sc > 0 and
LY
(c + a)(c + b) a−b
Sb ≥ 1 − = ≥ 0.
(c + a)2 c+a
O
If b2 Sa + a2 S b ≥ 0, then
.M
X a2
(b − c)2 Sa ≥ (b − c)2 Sa + (c − a)2 S b ≥ (b − c)2 Sa + 2 (b − c)2 S b
b
W
(b − c) (b Sa + a S b )
2 2 2
= ≥ 0.
b2
W
We have
W
2
a 2
2 2 2 2 b
b Sa + a S b = a + b − (a b + bc + ca) +
b+c c+a
2
2 2 b a 2
≥ a + b − (b + c)(c + a)) +
b+c c+a
b+c c+a

= a2 1 − + b2 1 −
c+a b+c
(a − b) (a b + bc + ca)
2
= ≥ 0.
(b + c)(c + a)
Second Solution. Multiplying by a b + bc + ca, the inequality becomes

X 2a2 b2 X 1 5
+ 2a bc + a2 + b2 + c 2 ≥ (a b + bc + ca),
(a + b)2 a+b 2
X 1 X1 X 4a b
2a bc + a2 + b2 + c 2 − 2(a b + bc + ca) − ab 1 − ≥ 0.
a+b 2 (a + b)2
According to the second solution of P 1.58-(a), we can write the inequality as follows
X a b(a − b)2 X a b(a − b)2
− ≥ 0,
(b + c)(c + a) 2(a + b)2
L
(b − c)2 Sa + (c − a)2 S b + (a − b)2 Sc ≥ 0,
.M
where
bc
D
Sa = [2(b + c)2 − (a + b)(a + c)].
b+c
A
Without loss of generality, assume that a ≥ b ≥ c. We have Sc > 0 and
Sb =
ac
[2(a + c)2 − (a + b)(b + c)] ≥
ac PI
[2(a + c)2 − (2a)(a + c)]
a+c a+c
M
2ac 2 (a + c)
= ≥ 0.
LY
a+c
If Sa + S b ≥ 0, then
O
X
(b − c)2 Sa ≥ (b − c)2 Sa + (a − c)2 S b ≥ (b − c)2 (Sa + S b ) ≥ 0.
.M
The inequality Sa + S b ≥ 0 is true if

W
ac bc
[2(a + c)2 − (a + b)(b + c)] ≥ [(a + b)(a + c) − 2(b + c)2 ].
a+c b+c
W
Since
ac bc
W
≥ ,
a+c b+c
2(a + c)2 − (a + b)(b + c) ≥ (a + b)(a + c) − 2(b + c)2 .
This is true since is equivalent to
(a − b)2 + 2c(a + b) + 4c 2 ≥ 0.
98 Vasile Cîrtoaje
ab bc ca 1 a b + bc + ca
+ + + ≥ 2 .
(a + b)2 (b + c)2 (c + a)2 4 a + b2 + c 2
First Solution. We use the SOS method. Write the inequality as follows
a b + bc + ca X 1

bc
1− ≥ − ,
a2 + b2 + c 2 4 (b + c)2
L
X (b − c)2 X (b − c)2
.M
2 ≥ ,
a2 + b2 + c 2 (b + c)2
a2 + b2 + c 2

D
X
2
(b − c) 2 − ≥ 0.
(b + c)2
A
Since
2−
a2 + b2 + c 2
= 1 +
2bc − a2
≥
PI
1 −
a 2
,
(b + c)2 (b + c)2 b+c
M
LY
(b − c)2 Sa + (c − a)2 S b + (a − b)2 Sc ≥ 0,

O
where 2
.M
a 2
b c 2
Sa = 1 − , Sb = 1 − , Sc = 1 − .
b+c c+a a+b
Without loss of generality, assume that a ≥ b ≥ c. Since S b ≥ 0 and Sc > 0, if b2 Sa +
W
a2 S b ≥ 0, then
W
X a2
(b − c)2 Sa ≥ (b − c)2 Sa + (c − a)2 S b ≥ (b − c)2 Sa + 2 (b − c)2 S b
W
b
(b − c)2 (b2 Sa + a2 S b )
= ≥ 0.
b2
We have
ab 2 ab 2

2 2 2 2
b Sa + a S b = a + b − −
b+c c+a
2 a 2
b
= a2 1 − + b2 1 − ≥ 0.
b+c c+a
Second Solution. Since (a + b)2 ≤ 2(a2 + b2 ), it suffices to prove that

X ab 1 a b + bc + ca
+ ≥ 2 ,
2(a2 +b ) 4
2 a + b2 + c 2
X 2a b 4(a b + bc + ca)
+1≥ ,
a +b
2 2 a2 + b2 + c 2
X (a + b)2 4(a b + bc + ca)
≥2+ ,
+
a2 b2 a2 + b2 + c 2
L
X (a + b)2 2(a + b + c)2
.M
≥ .
a2 + b2 a2 + b2 + c 2
The last inequality follows immediately by the Cauchy-Schwarz inequality
D
[ (a + b)]2
P
X (a + b)2
A
≥ P .
a2 + b2 (a2 + b2 )
PI
Remark. The following generalization of the inequalities in P 1.61 and P 1.62 holds:
M
• Let a, b, c be nonnegative real numbers, no two of which are zero. If 0 ≤ k ≤ 2, then
LY
X 4a b a2 + b2 + c 2 a b + bc + ca
+ k ≥ 3k − 1 + 2(2 − k) 2 .
(a + b) a b + bc + ca a + b2 + c 2
O
with equality for a = b = c, and for a = 0 and b = c (or any cyclic permutation).
.M
W
W
3a b 3bc 3ca a b + bc + ca 5
+ + ≤ 2 + .
W
(a + b)2 (b + c)2 (c + a)2 a + b2 + c 2 4
Solution. We use the SOS method. Write the inequality as follows

X1 a b + bc + ca

bc
3 − ≥1− 2 ,
4 (b + c) 2 a + b2 + c 2
X (b − c)2 X (b − c)2
3 ≥2 ,
(b + c)2 a2 + b2 + c 2
(b − c)2 Sa + (c − a)2 S b + (a − b)2 Sc ≥ 0,
100 Vasile Cîrtoaje
where
3(a2 + b2 + c 2 ) 3(a2 + b2 + c 2 ) 3(a2 + b2 + c 2 )

Sa = − 2, S b = − 2, S c = − 2.
(b + c)2 (c + a)2 (a + b)2
Without loss of generality, assume that a ≥ b ≥ c. Since Sa > 0 and
a2 + 3b2 + c 2 − 4ac (a − 2c)2 + 3(b2 − c 2 )

Sb = = > 0,
(c + a)2 (c + a)2
if S b + Sc ≥ 0, then
L
X
(b − c)2 Sa ≥ (c − a)2 S b + (a − b)2 Sc ≥ (a − b)2 (S b + Sc ) ≥ 0.
.M
Using the Cauchy-Schwarz Inequality, we have
D
1 1

S b + Sc = 3(a2 + b2 + c 2 ) + −4
A
(c + a)2 (a + b)2
≥
12(a2 + b2 + c 2 )
(c + a)2 + (a + b)2
− 4 =
PI
4(a − b − c)2 + 4(b − c)2
(c + a)2 + (a + b)2
≥ 0.
M
a
The equality occurs for a = b = c, and for = b = c (or any cyclic permutation).
LY
2
O
.M
a3 + a bc b3 + a bc c 3 + a bc
W
(a) + + ≥ a2 + b2 + c 2 ;
b+c c+a a+b
W
a3 + 2a bc b3 + 2a bc c 3 + 2a bc 1
(b) + + ≥ (a + b + c)2 ;
b+c c+a a+b
W
a3 + 3a bc b3 + 3a bc c 3 + 3a bc
(c) + + ≥ 2(a b + bc + ca).
b+c c+a a+b
Solution. (a) First Solution. Write the inequality as

X a3 + a bc
2
− a ≥ 0,
b+c
X a(a − b)(a − c)
≥ 0.
b+c
Assume that a ≥ b ≥ c. Since (c − a)(c − b) ≥ 0 and
a(a − b)(a − c) b(b − c)(b − a) (a − b)2 (a2 + b2 + c 2 + a b)

+ = ≥ 0,
b+c b+c (b + c)(c + a)
the conclusion follows. The equality occurs for a = b = c, and for a = 0 and b = c (or
(b) Taking into account the inequality in (a), it suffices to show that
a bc a bc a bc 1
+ + + a2 + b2 + c 2 ≥ (a + b + c)2 ,
b+c c+a a+b 2
L
which is just the inequality (a) from P 1.58. The equality occurs for a = b = c, and for
.M
a = 0 and b = c (or any cyclic permutation).
(c) The desired inequality follows by adding the inequality in (a) to the inequality (a)
D
from P 1.58. The equality occurs for a = b = c, and for a = 0 and b = c (or any cyclic
A
permutation).
PI
M
LY
a3 + 3a bc b3 + 3a bc c 3 + 3a bc
O
+ + ≥ a + b + c.
(b + c)2 (c + a)2 (a + b)2
.M
Solution. We use the SOS method. We have

W
X a3 + 3a bc X X a3 + 3a bc a − a(b2 − bc + c 2 )
X 3
− = − =
W
a a
(b + c)2 (b + c)2 (b + c)2
W
X a3 (b + c) − a(b3 + c 3 ) X a b(a2 − b2 ) + ac(a2 − c 2 )

= =
(b + c)3 (b + c)3
X a b(a2 − b2 ) X ba(b2 − a2 ) X a b(a2 − b2 )[(c + a)3 − (b + c)3 ]
= + =
(b + c)3 (c + a)3 (b + c)3 (c + a)3
X a b(a + b)(a − b)2 [(c + a)2 + (c + a)(b + c) + (b + c)2 ]
= ≥ 0.
(b + c)3 (c + a)3
a3 + 3a bc b3 + 3a bc c 3 + 3a bc 3
(a) + + ≥ ;
(b + c)3 (c + a)3 (a + b) 3 2
3a3 + 13a bc 3b3 + 13a bc 3c 3 + 13a bc

(b) + + ≥ 6.
(b + c)3 (c + a)3 (a + b)3
(Vasile Cîrtoaje and Ji Chen, 2005)
Solution. (a) First Solution. Use the SOS method. We have
L
X a3 + 3a bc X a(b + c)2 + a(a2 + bc − b2 − c 2 )
=
.M
(b + c)3 (b + c)3
X a X a3 − a(b2 − bc + c 2 )
= +
D
b+c (b + c)3
A
3 X a3 (b + c) − a(b3 + c 3 )
≥ +
2
3
(b + c)4 PI
X a b(a2 − b2 ) + ac(a2 − c 2 )
= +
M
2 (b + c)4
3 X a b(a2 − b2 ) X ba(b2 − a2 )
LY
= + +
2 (b + c)4 (c + a)4
X a b(a + b)(a − b)[(c + a)4 − (b + c)4 ]
O
3
= + ≥ 0.
2 (b + c)4 (c + a)4
.M
The equality occurs for a = b = c.

Second Solution. Assume that a ≥ b ≥ c. Since
W
a3 + 3a bc b3 + 3a bc c 3 + 3a bc
W
≥ ≥
b+c c + ac a+b
W
and
1 1 1
≥ ≥ ,
(b + c)2 (c + a)2 (a + b)2
by Chebyshev’s inequality, we get
X a3 + 3a bc 1 X a3 + 3a bc X

1
≥ .
(b + c)3 3 b+c (b + c)2
a3 + 3a bc X
X
1 9
≥ .
b+c (b + c)2 2
We can obtain this inequality by multiplying the known inequality (Iran-1996)

X 1 9
≥
(b + c)2 4(a b + bc + ca)
and the inequality (c) from P 1.64.
(b) We have
X 3a3 + 13a bc X 3a(b + c)2 + 4a bc + 3a(a2 + bc − b2 − c 2 )
=
(b + c)3 (b + c)3
X 3a X 1 X a3 − a(b2 − bc + c 2 )
= + 4a bc + 3 .
L
b+c (b + c)3 (b + c)3
.M
Since X 1 3
≥
(b + c)3 (a + b)(b + c)(c + a)
D
and
A
X a3 − a(b2 − bc + c 2 ) X a3 (b + c) − a(b3 + c 3 )
=
(b + c)3 PI
(b + c)4
X a b(a2 − b2 ) + ac(a2 − c 2 ) X a b(a2 − b2 ) X ba(b2 − a2 )
= = +
M
(b + c)4 (b + c)4 (c + a)4
LY
X a b(a + b)(a − b)[(c + a)4 − (b + c)4 ]

= ≥ 0,
(b + c)4 (c + a)4
O

.M
X 3a 12a bc
+ ≥ 6.
b + c (a + b)(b + c)(c + a)
W
This inequality is equivalent to the third degree Schur’s inequality

X
a3 + b3 + c 3 + 3a bc ≥ a b(a + b).
W
W
a3 b3 c3 3
(a) + + + a b + bc + ca ≥ (a2 + b2 + c 2 );
b+c c+a a+b 2
2a2 + bc 2b2 + ca 2c 2 + a b 9(a2 + b2 + c 2 )

(b) + + ≥ .
b+c c+a a+b 2(a + b + c)

Solution. (a) We apply the SOS method. Write the inequality as

X 2a3 X
2
−a ≥ (a − b)2 .
b+c
Since
X 2a3 a (a − b) + a2 (a − c)
X 2
2
−a =
b+c b+c
X a2 (a − b) X b2 (b − a) X (a − b)2 (a2 + b2 + a b + bc + ca)
= + = ,
b+c c+a (b + c)(c + a)
L
.M
(b − c)2 Sa + (c − a)2 S b + (a − b)2 Sc ≥ 0,
D
where
A
Sa = (b + c)(b2 + c 2 − a2 ), S b = (c + a)(c 2 + a2 − b2 ), Sc = (a + b)(a2 + b2 − c 2 ).
PI
Without loss of generality, assume that a ≥ b ≥ c. Since S b ≥ 0, Sc ≥ 0, and
M
Sa + S b = (a + b)(a − b)2 + c 2 (a + b + 2c) ≥ 0,
LY
we have
O
X
(b − c)2 Sa ≥ (b − c)2 Sa + (a − c)2 S b ≥ (b − c)2 (Sa + S b ) ≥ 0.
.M
(b) Multiplying by a + b + c, the inequality can be written as
W
X a 9
1+ (2a2 + bc) ≥ (a2 + b2 + c 2 ),
W
b+c 2
or
W
X 2a3 + a bc 5
+ a b + bc + ca ≥ (a2 + b2 + c 2 ).
b+c 2
This inequality follows using the inequality in (a) and the first inequality from P 1.58.
a(b + c) b(c + a) c(a + b)

+ 2 + 2 ≥ 2.
b2 + bc + c 2 c + ca + a 2 a + a b + b2
First Solution. Apply the SOS method. We have
a(b + c) a(b + c)(a + b + c)

X X
(a + b + c) −2 = − 2a
b2 + bc + c 2 b2 + bc + c 2
X a(a b + ac − b2 − c 2 ) X a b(a − b) − ca(c − a)
= =
b2 + bc + c 2 b2 + bc + c 2
X a b(a − b) X a b(a − b)
= −
b2 + bc + c 2 c 2 + ca + a2
X a b(a − b)2
= (a + b + c) ≥ 0.
(b2 + bc + c 2 )(c 2 + ca + a2 )
L
.M
Second Solution. By the AM-GM inequality, we have
D
4(b2 + bc + c 2 )(a b + bc + ca) ≤ (b2 + bc + c 2 + a b + bc + ca)2 = (b + c)2 (a + b + c)2 .
A
Thus,
X a(b + c)
b + bc + c
2 2
=
X PI
a(b + c)(a b + bc + ca)
(b + bc + c 2 )(a b + bc + ca)
2
M
X 4a(a b + bc + ca) 4(a b + bc + ca) X a
≥ = ,
LY
(b + c)(a + b + c)2 (a + b + c)2 b+c

O
X a (a + b + c)2
≥ .
.M
b+c 2(a b + bc + ca)

This follows immediately from the Cauchy-Schwarz inequality
W
X a (a + b + c)2
≥ P .
b+c a(b + c)
W
Third Solution. By the Cauchy-Schwarz inequality, we have

W
X a(b + c) (a + b + c)2
≥ .
b2 + bc + c 2 P a(b2 + bc + c 2 )
b+c
Thus, it is enough to show that
X a(b2 + bc + c 2 )
(a + b + c)2 ≥ 2 .
b+c
Since
a(b2 + bc + c 2 )

bc a bc
=a b+c− = a b + ca − ,
b+c b+c b+c
this inequality is equivalent to
1 1 1

2a bc + + + a2 + b2 + c 2 ≥ 2(a b + bc + ca),
b+c c+a a+b
which is just the inequality (a) from P 1.58.

Fourth Solution. By direct calculation, we can write the inequality as
X X
a b(a4 + b4 ) ≥ a2 b2 (a2 + b2 ),
L
X
a b(a − b)(a3 − b3 ) ≥ 0.
D
.M
A
PI
a(b + c) b(c + a) c(a + b) Y a − b 2
M
+ + ≥2+4 .
b2 + bc + c 2 c 2 + ca + a2 a2 + a b + b2 a+b
LY

O
Solution. For b = c = 1, the inequality reduces to a(a − 1)2 ≥ 0. Assume further that
a < b < c. According to the first solution of P 1.68, we have
.M
X a(b + c) X bc(b − c)2

− 2 = .
b2 + bc + c 2 (a2 + a b + b2 )(a2 + ac + c 2 )
W
Therefore, it remains to show that

W
X bc(b − c)2 Y a − b 2
W
≥4 .
(a2 + a b + b2 )(a2 + ac + c 2 ) a+b
Since
(a2 + a b + b2 )(a2 + ac + c 2 ) ≤ (a + b)2 (a + c)2 ,
X bc(b − c)2 Y a − b 2
≥4 ,
(a + b)2 (a + c)2 a+b
X bc(b + c)2
≥ 4.
(a − b)2 (a − c)2
We have
X bc(b + c)2 bc(b + c)2
≥
(a − b)2 (a − c)2 (a − b)2 (a − c)2
bc(b + c)2 (b + c)2
≥ = ≥ 4.
b2 c 2 bc
L
.M
a b − bc + ca bc − ca + a b ca − a b + bc 3
+ + ≥ .
b2 + c 2 c 2 + a2 a2 + b2 2
D
A
Solution. Use the SOS method. We have
X a b − bc + ca 1 X (b + c)(2a − b − c)
− =
PI
M
b2 + c 2 2 2(b2 + c 2 )
LY
X (b + c)(a − b) X (b + c)(a − c)
= +
2(b2 + c 2 ) 2(b2 + c 2 )
O
X (b + c)(a − b) X (c + a)(b − a)
= +
2(b2 + c 2 ) 2(c 2 + a2 )
.M
X (a − b)2 (a b + bc + ca − c 2 )
= .
2(b2 + c 2 )(c 2 + a2 )
W
Since
a b + bc + ca − c 2 = (b − c)(c − a) + 2a b ≥ (b − c)(c − a),
W

W
X
(a2 + b2 )(a − b)2 (b − c)(c − a) ≥ 0.

X
(a − b)(b − c)(c − a) (a − b)(a2 + b2 ) ≥ 0,
or
(a − b)2 (b − c)2 (c − a)2 ≥ 0.
P 1.71. Let a, b, c be nonnegative real numbers, no two of which are zero. If k > −2, then
X a b + (k − 1)bc + ca 3(k + 1)
≥ .
b2 + k bc + c2 k+2
First Solution. Apply the SOS method. Write the inequality as

X a b + (k − 1)bc + ca k+1

− ≥ 0,
b2 + k bc + c 2 k+2
L
X A
≥ 0,
.M
b2 + k bc + c 2
where
A = (b + c)(2a − b − c) + k(a b + ac − b2 − c 2 ).
D
A
Since
PI
A =(b + c)[(a − b) + (a − c)] + k[b(a − b) + c(a − c)]
M
= (a − b)[(k + 1)b + c] + (a − c)[(k + 1)c + b],
LY
the inequality is equivalent to

X (a − b)[(k + 1)b + c] X (a − c)[(k + 1)c + b]
O
+ ≥ 0,
b2 + k bc + c 2 b2 + k bc + c 2
.M
X (a − b)[(k + 1)b + c] X (b − a)[(k + 1)a + c]

+ ≥ 0,
b2 + k bc + c 2 c 2 + kca + a2
W
X
(b − c)2 R a Sa ≥ 0,
W
where
R a = b2 + k bc + c 2 , Sa = a(b + c − a) + (k + 1)bc.
W
Without loss of generality, assume that a ≥ b ≥ c.
Case 1: k ≥ −1. Since Sa ≥ a(b + c − a), it suffices to show that

X
a(b + c − a)(b − c)2 R a ≥ 0.
We have
X
a(b + c − a)(b − c)2 R a ≥ a(b + c − a)(b − c)2 R a + b(c + a − b)(c − a)2 R b
≥ (b − c)2 [a(b + c − a)R a + b(c + a − b)R b ].
Thus, it is enough to prove that
a(b + c − a)R a + b(c + a − b)R b ≥ 0.
Since b + c − a ≥ −(c + a − b), we have
a(b + c − a)R a + b(c + a − b)R b ≥ (c + a − b)(bR b − aR a )

= (c + a − b)(a − b)(a b − c 2 ) ≥ 0.
Case 2: −2 < k ≤ 1. Since
Sa = (a − b)(c − a) + (k + 2)bc ≥ (a − b)(c − a),
L
.M
we have X X
(b − c)2 R a Sa ≥ (a − b)(b − c)(c − a) (b − c)R a .
D
From
A
X X
(b − c)R a = (b − c)[b2 + bc + c 2 − (1 − k)bc]
=
X
(b3 − c 3 ) − (1 − k)
X PI
bc(b − c)
M
= (1 − k)(a − b)(b − c)(c − a),
LY
we get
O
X
(a − b)(b − c)(c − a) (b − c)R a = (1 − k)(a − b)2 (b − c)2 (c − a)2 ≥ 0.
.M
This completes the proof. The equality occurs for a = b = c, and for a = 0 and b = c
(or any cyclic permutation).
Second Solution. Let p = a + b + c, q = a b + bc + ca, r = a bc. Write the inequality in
W
the form f6 (a, b, c) ≥ 0, where

W
X
f6 (a, b, c) = (k + 2) [a(b + c) + (k − 1)bc](a2 + ka b + b2 )(a2 + kac + c 2 )
W
Y
−3(k + 1) (b2 + k bc + c 2 )
X
= (k + 2) [(k − 2)bc + q](ka b − c 2 + p2 − 2q)(kac − b2 + p2 − 2q)
Y
−3(k + 1) (k bc − a2 + p2 − 2q).
Thus, f6 (a, b, c) has the same highest coefficient A as
(k + 2)(k − 2)P2 (a, b, c) − 3(k + 1)P3 (a, b, c),
where X
P2 (a, b, c) = bc(ka b − c 2 )(kac − b2 ),
Y
P3 (a, b, c) = (k bc − a2 ).
According to Remark 2 from the proof of P 2.75 in Volume 1,
A = (k + 2)(k − 2)P2 (1, 1, 1) − 3(k + 1)P3 (1, 1, 1)

= 3(k + 2)(k − 2)(k − 1)2 − 3(k + 1)(k − 1)3 = −9(k − 1)2 .
Since A ≤ 0, according to P 3.76-(a) in Volume 1, it suffices to prove the original in-

equality for b = c = 1, and for a = 0. For these cases, this inequality has the obvious
forms
(k + 2)a(a − 1)2 ≥ 0
L
and
.M
(b − c)2 [(k + 2)(b2 + c 2 ) + (k2 + k + 1)bc] ≥ 0,
respectively.
D
Remark. For k = 1 and k = 0, from P 1.71, we get the inequalities in P 1.68 and P 1.70,
A
respectively. Besides, for k = 2, we get the well-known inequality (Iran 1996):
1
+
1
+
1
≥
9
PI .
(a + b) (b + c) (c + a) 4(a b + bc + ca)
M
2 2 2
LY
O
.M
X 3bc − a(b + c) 3
≤ .
b2 + k bc + c 2 k+2
W

W
Solution. Write the inequality in P 1.71 as

W
X a b + (k − 1)bc + ca 3

1− ≤ ,
b + k bc + c
2 2 k+2
X b2 + c 2 + bc − a(b + c) 3
≤ .
b2 + k bc + c 2 k+2
Since b2 + c 2 ≥ 2bc, we get
X 3bc − a(b + c) 3
≤ ,
b2 + k bc + c2 k+2
which is just the desired inequality. The equality occurs for a = b = c.
ab + 1 bc + 1 ca + 1 4
+ 2 + 2 ≥ .
a +b
2 2 b +c 2 c +a 2 3
Solution. Write the inequality in the homogeneous form E(a, b, c) ≥ 4, where
4a b + bc + ca 4bc + ca + a b 4ca + a b + bc
E(a, b, c) = + + .
a2 + b2 b2 + c 2 c 2 + a2
Without loss of generality, assume that a = min{a, b, c}. We will show that
L
E(a, b, c) ≥ E(0, b, c) ≥ 4.
.M
For a = 0, we have E(a, b, c) = E(0, b, c), and for a > 0, we have
D
E(a, b, c) − E(0, b, c) 4b2 + c(b − a) b+c 4c 2 + b(c − a)
= + + > 0.
b(a2 + b2 ) b2 + c 2 c(c 2 + a2 )
A
a
Also,
b 4bc c
PI
(b − c)4
E(0, b, c) − 4 = + 2 + − 4 = ≥ 0.
M
c b + c2 b bc(b2 + c 2 )
p
LY
The equality holds for a = 0 and b = c = 3 (or any cyclic permutation).

O
.M
5a b + 1 5bc + 1 5ca + 1
+ + ≥ 2.
W
(a + b)2 (b + c)2 (c + a)2

W
Solution. Write the inequality as E(a, b, c) ≥ 6, where

W
16a b + bc + ca 16bc + ca + a b 16ca + a b + bc

E(a, b, c) = + + .
(a + b)2 (b + c)2 (c + a)2
Without loss of generality, assume that a ≤ b ≤ c.
Case 1: 16b2 ≥ c(a + b). We will show that
E(a, b, c) ≥ E(0, b, c) ≥ 6.
For a = 0, we have E(a, b, c) = E(0, b, c), and for a > 0, we have
E(a, b, c) − E(0, b, c) 16b2 − c(a + b) 1 16c 2 − b(a + c)

= + + > 0.
a b(a + b)2 b+c c(c + a)2
Also,
b 16bc c (b − c)4
E(0, b, c) − 6 = + + − 6 = ≥ 0.
c (b + c)2 b bc(b + c)2
Case 2: 16b2 < c(a + b). We have
16a b + bc + ca 16a b + 16b2 2(5b − 3a)
E(a, b, c) − 6 > − 6 > −6= > 0.
(a + b)2 (a + b)2 a+b
p
L
.M
a2 − bc b2 − ca c2 − a b
+ + ≥ 0.
2b2 − 3bc + 2c 2 2c 2 − 3ca + 2a2 2a2 − 3a b + 2b2
D
A
Solution. The hint is applying the Cauchy-Schwarz inequality after we made the nu-
PI
merators of the fractions to be nonnegative and as small as possible. Thus, we write the
inequality as
M
X a2 − bc

+ 1 ≥ 3,
2b2 − 3bc + 2c 2
LY
X a2 + 2(b − c)2
≥ 3.
O
2b2 − 3bc + 2c 2
Using the Cauchy-Schwarz inequality
.M
(5 a2 − 4 a b)2
P P
X a2 + 2(b − c)2
≥P ,
2b2 − 3bc + 2c 2 [a2 + 2(b − c)2 ](2b2 − 3bc + 2c 2 )
W

W
X X X
(5 a2 − 4 a b)2 ≥ 3 [a2 + 2(b − c)2 ](2b2 − 3bc + 2c 2 ).
W

X X X X
a4 + a bc a+2 a b(a2 + b2 ) ≥ 6 a2 b2 .
We can obtain it by summing Schur’s inequality of degree four
X X X
a4 + a bc a≥ a b(a2 + b2 )
to the obvious inequality
X X
3 a b(a2 + b2 ) ≥ 6 a2 b2 .
2a2 − bc 2b2 − ca 2c 2 − a b
+ + ≥ 3.
b2 − bc + c 2 c 2 − ca + a2 a2 − a b + b2
Solution. Write the inequality such that the numerators of the fractions are nonnegative
and as small as possible:
X 2a2 − bc
+ 1 ≥ 6,
b2 − bc + c 2
L
X 2a2 + (b − c)2
≥ 6.
.M
b2 − bc + c 2
Applying the Cauchy-Schwarz inequality, we get
D
4(2 a2 − a b)2
P P
X 2a2 + (b − c)2
A
≥P .
b2 − bc + c 2 [2a2 + (b − c)2 ](b2 − bc + c 2 )
Thus, we still have to prove that

PI
M
X X X
2(2 a2 − a b)2 ≥ 3 [2a2 + (b − c)2 ](b2 − bc + c 2 ).
LY

O
X X X X
2 a4 + 2a bc a+ a b(a2 + b2 ) ≥ 6 a2 b2 .
.M
We can obtain it by summing up Schur’s inequality of degree four

X X X
W
a4 + a bc a≥ a b(a2 + b2 )
W
and X X
a b(a2 + b2 ) ≥ 2 a2 b2 ,
W
multiplied by 2 and 3, respectively. The equality holds for a = b = c, and for a = 0 and
b = c (or any cyclic permutation).
a2 b2 c2
+ + ≥ 1.
2b2 − bc + 2c 2 2c 2 − ca + 2a2 2a2 − a b + 2b2
( a2 )2
P
X a2
≥P .
2b2 − bc + 2c 2 a2 (2b2 − bc + 2c 2 )
X X
( a2 )2 ≥ a2 (2b2 − bc + 2c 2 ),
which is equivalent to X X X
a4 + a bc a≥2 a2 b2 .
This inequality follows by adding Schur’s inequality of degree four
L
.M
X X X
a4 + a bc a≥ a b(a2 + b2 )
and
D
X X
a b(a2 + b2 ) ≥ 2 a2 b2 .
A
PI
M
LY
1 1 1 9
+ 2 + 2 ≥ .
− bc + 4c 4c − ca + 4a 4a − a b + 4b 7(a + b2 + c 2 )
O
4b2 2 2 2 2

.M
Solution. We use the SOS method. Without loss of generality, assume that a ≥ b ≥ c.
Write the inequality as
W
X 7(a2 + b2 + c 2 )
− 3 ≥ 0,
4b2 − bc + 4c 2
W
X 7a2 − 5b2 − 5c 2 + 3bc

≥ 0,
W
4b2 − bc + 4c 2
X 5(2a2 − b2 − c 2 ) − 3(a2 − bc)
≥ 0.
4b2 − bc + 4c 2
Since
2(a2 − bc) = (a − b)(a + c) + (a − c)(a + b),
we have
10(2a2 − b2 − c 2 ) − 6(a2 − bc) =

= 10(a2 − b2 ) − 3(a − b)(a + c) + 10(a2 − c 2 ) − 3(a − c)(a + b)
= (a − b)(7a + 10b − 3c) + (a − c)(7a + 10c − 3b).
Thus, we can write the desired inequality as follows

X (a − b)(7a + 10b − 3c) X (a − c)(7a + 10c − 3b)
+ ≥ 0,
4b2 − bc + 4c 2 4b2 − bc + 4c 2
X (a − b)(7a + 10b − 3c) X (b − a)(7b + 10a − 3c)
+ ≥ 0,
4b2 − bc + 4c 2 4c 2 − ca + 4a2
X (a − b)2 (28a2 + 28b2 − 9c 2 + 68a b − 19bc − 19ca)
,
(4b2 − bc + 4c 2 )(4c 2 − ca + 4a2 )
(b − c)2 Sa + (c − a)2 S b + (a − b)2 Sc ≥ 0,
L
.M
where
Sa = (4b2 − bc + 4c 2 )[(b − a)(28b + 9a) + c(−19a + 68b + 28c)],
D
A
S b = (4c 2 − ca + 4a2 )[(a − b)(28a + 9b) + c(−19b + 68a + 28c)],
PI
Sc = (4a2 − a b + 4b2 )[(b − c)(28b + 9c) + a(68b − 19c + 28a)].
M
Since S b ≥ 0 and Sc > 0, it suffices to show that Sa + S b ≥ 0. We have
LY
Sa ≥ (4b2 − bc + 4c 2 )[(b − a)(28b + 9a) − 19ac],

O
S b ≥ (4c 2 − ca + 4a2 )[(a − b)(28a + 9b) + 19ac],

.M
19ac[(4c 2 − ca + 4a2 ) − (4b2 − bc + 4c 2 ] = 19ac(a − b)(4a + 4b − c) ≥ 0,
and hence
W
Sa + S b ≥ (a − b)[−(4b2 − bc + 4c 2 )(28b + 9a) + (4c 2 − ca + 4a2 )(28a + 9b)]

W
= (a − b)2 [112(a2 + a b + b2 ) + 76c 2 − 28c(a + b)] ≥ 0.

W
2a2 + bc 2b2 + ca 2c 2 + a b 9
+ + ≥ .
b2 + c 2 c 2 + a2 a2 + b2 2

First Solution. We apply the SOS method. Since

X 2(2a2 + bc) X 2a2 − b2 − c 2 X (b − c)2
− 3 = 2 −
b2 + c 2 b2 + c 2 b2 + c 2
and
X 2a2 − b2 − c 2 X a2 − b2 X a2 − c 2 X a2 − b2 X b2 − a2
= + = +
b2 + c 2 b2 + c 2
b2 + c 2 b2 + c 2 c 2 + a2
1 1 (a2 − b2 )2
X X
= (a2 − b2 ) 2 − =
b + c 2 c 2 + a2 (b2 + c 2 )(c 2 + a2 )
L
.M
X (a − b)2 (a2 + b2 )
≥
(b2 + c 2 )(c 2 + a2 )
D
A
X (b − c)2 (b2 + c 2 ) X (b − c)2
2 ≥ PI ,
(c 2 + a2 )(a2 + b2 ) b2 + c 2
M
or
(b − c)2 Sa + (c − a)2 S b + (c − a)2 Sc ≥ 0,
LY
where
Sa = 2(b2 + c 2 )2 − (c 2 + a2 )(a2 + b2 ).
O
Assume that a ≥ b ≥ c. We have

.M
S b = 2(c 2 + a2 )2 − (a2 + b2 )(b2 + c 2 )

≥ 2(c 2 + a2 )(c 2 + b2 ) − (a2 + b2 )(b2 + c 2 )
W
= (b2 + c 2 )(a2 − b2 + 2c 2 ) ≥ 0,
W
Sc = 2(a2 + b2 )2 − (b2 + c 2 )(c 2 + a2 ) > 0,

W
Sa + S b = (a2 − b2 )2 + 2c 2 (a2 + b2 + 2c 2 ) ≥ 0.
Therefore,
(b − c)2 Sa + (c − a)2 S b + (c − a)2 Sc ≥ (b − c)2 Sa + (c − a)2 S b

≥ (b − c)2 (Sa + S b ) ≥ 0.

2b2 c 2
bc ≥ ,
b2 + c 2
we have
X 2a2 + bc X 2a2 + 2b2 c 2

b2 +c 2
X 1
≥ = 2(a2 b2 + b2 c 2 + c 2 a2 ) .
b2 + c 2 b2 + c 2 (b2 + c 2 )2

X 1 9
≥ ,
(b2 + c 2 )2 4(a2 b2 + b2 c 2 + c 2 a2 )
which is just the known Iran-1996 inequality (see Remark from P 1.71).
L
.M
Third Solution. We get the desired inequality by summing the inequality in P 1.59-(a),
namely
2a2 − 2bc 2b2 − 2ca 2c 2 − 2a b 6(a b + bc + ca)
D
+ 2 + 2 + ≥ 6,
b2 + c 2 c + a2 a + b2 a2 + b2 + c 2
A
and the inequality
3bc 3ca 3a b
PI
3 6(a b + bc + ca)
+ 2 + 2 + ≥ .
M
b2 +c 2 c +a 2 a +b 2 2 a2 + b2 + c 2
LY

X 2bc 4(a b + bc + ca)

O
+1 ≥ + 2,
b +c
2 2 a2 + b2 + c 2
.M
X (b + c)2 2(a + b + c)2

≥ .
b2 + c 2 a2 + b2 + c 2
W

W
P 2
X (b + c)2 (b + c) 2(a + b + c)2
≥ P = 2 .
W
b2 + c 2 (b2 + c 2 ) a + b2 + c 2
2a2 + 3bc 2b2 + 3ca 2c 2 + 3a b

+ + ≥ 5.
b2 + bc + c 2 c 2 + ca + a2 a2 + a b + b2


X 3(2a2 + 3bc)
− 5 ≥ 0,
b2 + bc + c 2
or
X 6a2 + 4bc − 5b2 − 5c 2
≥ 0.
b2 + bc + c 2
Since
2(a2 − bc) = (a − b)(a + c) + (a − c)(a + b),
we have
L
6a2 + 4bc − 5b2 − 5c 2 = 5(2a2 − b2 − c 2 ) − 4(a2 − bc)
.M
= 5(a2 − b2 ) − 2(a − b)(a + c) + 5(a2 − c 2 ) − 2(a − c)(a + b)
D
= (a − b)(3a + 5b − 2c) + (a − c)(3a + 5c − 2b).
A
Thus, we can write the desired inequality as follows
X (a − b)(3a + 5b − 2c)
+
PI
X (a − c)(3a + 5c − 2b)
≥ 0,
b2 + bc + c 2 b2 + bc + c 2
M
X (a − b)(3a + 5b − 2c) X (b − a)(3b + 5a − 2c)
LY
+ ≥ 0,
b2 + bc + c 2 c 2 + ca + a2
O
X (a − b)2 (3a2 + 3b2 − 4c 2 + 8a b + bc + ca)

,
(b2 + bc + c 2 )(c 2 + ca + a2 )
.M
(b − c)2 Sa + (c − a)2 S b + (a − b)2 Sc ≥ 0,

where
W
Sa = (b2 + bc + c 2 )(3b2 + 3c 2 − 4a2 + a b + 8bc + ca).

W
Assume that a ≥ b ≥ c. Since

W
S b = (c 2 + ca + a2 )[(a − b)(3a + 4b) + c(8a + b + c)] ≥ 0
and
Sc = (a2 + a b + b2 )[(b − c)(3b + 4c) + a(8b + c + a)] > 0,
it suffices to show that Sa + S b ≥ 0. We have
Sa + S b ≥ (b2 + bc + c 2 )(b − a)(3b + 4a) + (c 2 + ca + a2 )(a − b)(3a + 4b)

= (a − b)2 [3(a + b)(a + b + c) + a b − c 2 ] ≥ 0.
2a2 + 5bc 2b2 + 5ca 2c 2 + 5a b 21
+ + ≥ .
(b + c)2 (c + a)2 (a + b)2 4
Solution. Use the SOS method.Write the inequality as follows
X 2a2 + 5bc 7
− ≥ 0,
(b + c)2 4
X 4(a2 − b2 ) + 4(a2 − c 2 ) − 3(b − c)2
≥ 0,
L
(b + c)2
.M
X b2 − c 2 X c 2 − b2 X (b − c)2
4 + 4 − 3 ≥ 0,
(c + a)2 (a + b)2 (b + c)2
D
X (b − c)2 (b + c)(2a + b + c) X (b − c)2
4 − 3 ≥ 0.
(c + a)2 (a + b)2 (b + c)2
A
Substituting b + c = x, c + a = y and a + b = z, we can rewrite the inequality in the
form
PI
( y − z)2 S x + (z − x)2 S y + (x − y)2 Sz ≥ 0,
M
where
LY
S x = 4x 3 ( y + z) − 3 y 2 z 2 , S y = 4 y 3 (z + x) − 3z 2 x 2 , Sz = 4z 3 (x + y) − 3x 2 y 2 .
O
Without loss of generality, assume that 0 < x ≤ y ≤ z, z ≤ x + y. We have Sz > 0 and

.M
S y ≥ 4x 2 y(z + x) − 3x 2 z(x + y) = x 2 [4x y + z( y − 3x)] ≥ 0,

since for the nontrivial case y − 3x < 0, we get
W
4x y + z( y − 3x) ≥ 4x y + (x + y)( y − 3x) = x 2 (3x + y)( y − x) ≥ 0.

W
Thus, it suffices to show that S x + S y ≥ 0. Since
S x + S y = 4x y(x 2 + y 2 ) + 4(x 3 + y 3 )z − 3(x 2 + y 2 )z 2

W
≥ 4x y(x 2 + y 2 ) + 4(x 3 + y 3 )z − 3(x 2 + y 2 )(x + y)z

= 4x y(x 2 + y 2 ) + (x 2 − 4x y + y 2 )(x + y)z,
for the nontrivial case x 2 − 4x y + y 2 < 0, we get
S x + S y ≥ 4x y(x 2 + y 2 ) + (x 2 − 4x y + y 2 )(x + y)2
≥ 2x y(x + y)2 + (x 2 − 4x y + y 2 )(x + y)2
= (x − y)2 (x + y)2 .
X 2a2 + (2k + 1)bc 3(2k + 3)
≥ .
b2 + k bc + c 2 k+2
First Solution. There are two cases to consider.

Case 1: −2 < k ≤ −1/2. Write the inequality as
X 2a2 + (2k + 1)bc 2k + 1 6
− ≥ ,
b + k bc + c
2 2 k+2 k+2
L
.M
X 2(k + 2)a2 − (2k + 1)(b − c)2
≥ 6.
b2 + k bc + c 2
D
Since 2(k + 2)a2 − (2k + 1)(b − c)2 ≥ 0 for −2 < k ≤ −1/2, we can apply the Cauchy-
A
Schwarz inequality. Thus, it suffices to show that
P P PI
[2(k + 2) a2 − (2k + 1) (b − c)2 ]2
≥ 6,
M
[2(k + 2)a2 − (2k + 1)(b − c)2 ](b2 + k bc + c 2 )
P
LY
which is equivalent to each of the following inequalities
2[(1 − k) a2 + (2k + 1) a b]2

P P
O
≥ 3,
[2(k + 2)a2 − (2k + 1)(b − c)2 ](b2 + k bc + c 2 )
P
.M
X X X X
2(k + 2) a4 + 2(k + 2)a bc a − (2k + 1) a b(a2 + b2 ) ≥ 6 a2 b2 ,
W
X X X X
2(k + 2)[ a4 + a bc a− a b(a2 + b2 )] + 3 a b(a − b)2 ≥ 0.
W
The last inequality is true since, by Schur’s inequality of degree four, we have
X X X
a4 + a bc a b(a2 + b2 ) ≥ 0.
W
a−
Case 2: k ≥ −9/5. Use the SOS method. Without loss of generality, assume that a ≥
b ≥ c. Write the inequality as
X 2a2 + (2k + 1)bc 2k + 3
− ≥ 0,
b2 + k bc + c 2 k+2
X 2(k + 2)a2 − (2k + 3)(b2 + c 2 ) + 2(k + 1)bc
≥ 0,
b2 + k bc + c 2
X (2k + 3)(2a2 − b2 − c 2 ) − 2(k + 1)(a2 − bc)
≥ 0.
b2 + k bc + c 2
Since
2(a2 − bc) = (a − b)(a + c) + (a − c)(a + b),
we have
(2k + 3)(2a2 − b2 − c 2 ) − 2(k + 1)(a2 − bc) =
= (2k + 3)(a2 − b2 ) − (k + 1)(a − b)(a + c) + (2k + 3)(a2 − c 2 ) − (k + 1)(a − c)(a + b)
= (a − b)[(k + 2)a + (2k + 3)b − (k + 1)c] + (a − c)[(k + 2)a + (2k + 3)c − (k + 1)b).
Thus, we can write the desired inequality as
X (a − b)[(k + 2)a + (2k + 3)b − (k + 1)c]
+
L
b2 + k bc + c 2
.M
X (a − c)[(k + 2)a + (2k + 3)c − (k + 1)b]
+ ≥ 0,
b2 + k bc + c 2
D
or X (a − b)[(k + 2)a + (2k + 3)b − (k + 1)c]
+
A
b2 + k bc + c 2
PI
X (b − a)[(k + 2)b + (2k + 3)a − (k + 1)c]
≥ 0,
c 2 + kca + a2
M
or
(b − c)2 R a Sa + (c − a)2 R b S b + (a − b)2 R c Sc ≥ 0,
LY
where
O
R a = b2 + k bc + c 2 , R b = c 2 + kca + a2 , R c = a2 + ka b + b2 ,
Sa = (k + 2)(b2 + c 2 ) − (k + 1)2 a2 + (3k + 5)bc + (k2 + k − 1)a(b + c)
.M
= −(a − b)[(k + 1)2 a + (k + 2)b] + c[(k2 + k − 1)a + (3k + 5)b + (k + 2)c],

S b = (k + 2)(c 2 + a2 ) − (k + 1)2 b2 + (3k + 5)ca + (k2 + k − 1)b(c + a)
W
= (a − b)[(k + 2)a + (k + 1)2 b] + c[(3k + 5)a + (k2 + k − 1)b + (k + 2)c],

W
Sc = (k + 2)(a2 + b2 ) − (k + 1)2 c 2 + (3k + 5)a b + (k2 + k − 1)c(a + b)

W
= (k + 2)(a2 + b2 ) + (3k + 5)a b + c[(k2 + k − 1)(a + b) − (k + 1)2 c]

≥ (5k + 9)a b + c[(k2 + k − 1)(a + b) − (k + 1)2 c].
We have S b ≥ 0, since for the nontrivial case
(3k + 5)a + (k2 + k − 1)b + (k + 2)c < 0,
we get
S b ≥ (a − b)[(k + 2)a + (k + 1)2 b] + b[(3k + 5)a + (k2 + k − 1)b + (k + 2)c]
= (k + 2)(a2 − b2 ) + (k + 2)2 a b + (k + 2)bc > 0.

Also, we have Sc ≥ 0 for k ≥ −9/5, since
(5k + 9)a b + c[(k2 + k − 1)(a + b) − (k + 1)2 c] ≥

≥ (5k + 9)ac + c[(k2 + k − 1)(a + b) − (k + 1)2 c]
= (k + 2)(k + 4)ac + (k2 + k − 1)bc − (k + 1)2 c 2
≥ (2k2 + 7k + 7)bc − (k + 1)2 c 2
≥ (k + 2)(k + 3)c 2 ≥ 0.
Therefore, it suffices to show that R a Sa + R b S b ≥ 0. Since
L
bR b − aR a = (a − b)(a b − c 2 ) ≥ 0
.M
and
a
R a Sa + R b S b ≥ R a (Sa + S b ),
D
b
A
a
Sa + S b ≥ 0. PI
b
We have
M
bSa + aS b = (k + 2)(a + b)(a − b)2 + c f (a, b, c)
LY
≥ 2(k + 2)b(a − b)2 + c f (a, b, c),

O
a c
Sa + S b ≥ 2(k + 2)(a − b)2 + f (a, b, c),
.M
b b
where
W
f (a, b, c) = b[(k2 + k − 1)a + (3k + 5)b] + a[(3k + 5)a + (k2 + k − 1)b]

W
+(k + 2)c(a + b) = (3k + 5)(a2 + b2 ) + 2(k2 + k − 1)a b + (k + 2)c(a + b).

For the nontrivial case f (a, b, c) < 0, we have
W
a
Sa + S b ≥ 2(k + 2)(a − b)2 + f (a, b, c)
b
≥ 2(k + 2)(a − b)2 + (3k + 5)(a2 + b2 ) + 2(k2 + k − 1)a b

= (5k + 9)(a2 + b2 ) + 2(k2 − k − 5)a b ≥ 2(k + 2)2 a b ≥ 0.
The proof is completed. The equality holds for a = b = c, and for a = 0 and b = c (or
Second Solution. Let
p = a + b + c, q = a b + bc + ca.
Write the inequality as f6 (a, b, c) ≥ 0, where

X
f6 (a, b, c) = (k + 2) [2a2 + (2k + 1)bc](a2 + ka b + b2 )(a2 + kac + c 2 )
Y
−3(2k + 3) (b2 + k bc + c 2 ).
Since
(a2 + ka b + b2 )(a2 + kac + c 2 ) = (p2 − 2q + ka b − c 2 )(p2 − 2q + kac − b2 ),
b2 + k bc + c 2 = p2 − 2q + k bc − a2 ,
L
f6 (a, b, c has the same highest coefficient A as
.M
(k + 2)P2 (a, b, c) − 3(2k + 3)P3 (a, b, c),
D
where
A
X Y
P2 (a, b, c) = [2a2 +(2k +1)bc](ka b − c 2 )(kac − b2 ), P3 (a, b, c) = (b2 + k bc + c 2 ).
PI
According to Remark 2 from the proof of P 2.75 in Volume 1, we have
M
A = (k + 2)P2 (1, 1, 1) − 3(2k + 3)P3 (1, 1, 1) = 9(2k + 3)(k − 1)2 .
LY
On the other hand,

O
f6 (a, 1, 1) = 2(k + 2)a(a2 + ka + 1)(a − 1)2 (a + k + 2) ≥ 0,

.M
f6 (0, b, c) = (b − c)2 2(k + 2)(b2 + c 2 )2 + 2(k + 2)2 bc(b2 + c 2 ) + (4k2 + 6k − 1)b2 c 2 .

W
For −2 < k ≤ −3/2, we have A ≤ 0. According to P 3.76-(a) in Volume 1, it suffices

to show that f6 (a, 1, 1) ≥ 0 and f6 (0, b, c) ≥ 0 for all a, b, c ≥ 0. The first condition is
W
clearly satisfied. The second condition is also satisfied since

W
2(k + 2)(b2 + c 2 )2 + (4k2 + 6k − 1)b2 c 2 ≥ [8(k + 2) + 4k2 + 6k − 1]b2 c 2

= (4k2 + 14k + 15)b2 c 2 ≥ 0.
For k > −3/2, when A ≥ 0, we will apply the highest coefficient cancellation method.
Consider two cases: p2 ≤ 4q and p2 > 4q.
Case 1: p2 ≤ 4q. Since

f6 (1, 1, 1) = f6 (0, 1, 1) = 0,
define the homogeneous function
P(a, b, c) = a bc + B(a + b + c)3 + C(a + b + c)(a b + bc + ca)

such that P(1, 1, 1) = P(0, 1, 1) = 0; that is,
1 4
P(a, b, c) = a bc + (a + b + c)3 − (a + b + c)(a b + bc + ca).
9 9
We will prove the sharper inequality g6 (a, b, c) ≥ 0, where
g6 (a, b, c) = f6 (a, b, c) − 9(2k + 3)(k − 1)2 P 2 (a, b, c).
Clearly, g6 (a, b, c) has the highest coefficient A = 0. Then, according to Remark 1 from
the proof of P 3.76 in Volume 1, it suffices to prove that g6 (a, 1, 1) ≥ 0 for 0 ≤ a ≤ 4.
We have
L
a(a − 1)2
P(a, 1, 1) = ,
.M
9
hence
D
a(a − 1)2 g(a)
A
g6 (a, 1, 1) = f6 (a, 1, 1) − 9(2k + 3)(k − 1)2 P 2 (a, 1, 1) = ,
9
where
PI
M
g(a) = 18(k + 2)(a2 + ka + 1)(a + k + 2) − (2k + 3)(k − 1)2 a(a − 1)2 .
LY
Since a2 + ka + 1 ≥ (k + 2)a, it suffices to show that

O
18(k + 2)2 (a + k + 2) ≥ (2k + 3)(k − 1)2 (a − 1)2 .

.M
Also, since (a − 1)2 ≤ 2a + 1, it is enough to prove that h(a) ≥ 0, where
h(a) = 18(k + 2)2 (a + k + 2) − (2k + 3)(k − 1)2 (2a + 1).

W
W
Since h(a) is a linear function, the inequality h(a) ≥ 0 is true if h(0) ≥ 0 and h(4) ≥ 0.
Setting x = 2k + 3, x > 0, we get
W
1
h(0) = 18(k + 2)3 − (2k + 3)(k − 1)2 = (8x 3 + 37x 2 + 2x + 9) > 0.
4
Also,
1
h(4) = 2(k + 2)2 (k + 6) − (2k + 3)(k − 1)2 = 3(7k2 + 20k + 15) > 0.
9
Case 2: p2 > 4q. We will prove the sharper inequality g6 (a, b, c) ≥ 0, where
g6 (a, b, c) = f6 (a, b, c) − 9(2k + 3)(k − 1)2 a2 b2 c 2 .

We see that g6 (a, b, c) has the highest coefficient A = 0. According to Remark 1 from
the proof of P 3.76 in Volume 1, it suffices to prove that g6 (a, 1, 1) ≥ 0 for a > 4 and
g6 (0, b, c) ≥ 0 for all b, c ≥ 0. We have
g6 (a, 1, 1) = f6 (a, 1, 1) − 9(2k + 3)(k − 1)2 a2
= a[2(k + 2)(a2 + ka + 1)(a − 1)2 (a + k + 2) − 9(2k + 3)(k − 1)2 a].

Since
a2 + ka + 1 > (k + 2)a, (a − 1)2 > 9,
L
.M
2(k + 2)2 (a + k + 2) ≥ (2k + 3)(k − 1)2 .
Indeed,
D
A
2(k + 2)2 (a + k + 2) − (2k + 3)(k − 1)2 > 2(k + 2)2 (k + 6) − (2k + 3)(k − 1)2
PI
= 3(7k2 + 20k + 15) > 0.
M
Also,
g6 (0, b, c) = f6 (0, b, c) ≥ 0.
LY
O
.M
3bc − 2a2 3
W
X
≤ .
b2 + k bc + c 2 k+2
W

W

X 2a2 − 3bc 3

6
+ ≥ ,
b2 + k bc + c 2 k + 2 k+2
X 2(k + 2)a2 + 3(b − c)2

≥ 6.
b2 + k bc + c 2
Applying the Cauchy-Schwarz inequality, it suffices to show that
[2(k + 2) a2 + 3 (b − c)2 ]2
P P
≥ 6,
[2(k + 2)a2 + 3(b − c)2 ](b2 + k bc + c 2 )
P
which is equivalent to each of the following inequalities
2[(k + 5) a2 − 3 a b]2
P P
≥ 3,
[2(k + 2)a2 + 3(b − c)2 ](b2 + k bc + c 2 )
P
X X X X
2(k + 8) a4 + 2(2k + 19) a2 b2 ≥ 6(k + 2)a bc a + 21 a b(a2 + b2 ),
2(k + 2) f (a, b, c) + 3g(a, b, c) ≥ 0,
where X X X
f (a, b, c) = a4 + 2 a2 b2 − 3a bc a,
X X X
g(a, b, c) = 4 a4 + 10 a2 b2 − 7 a b(a2 + b2 ).
L
.M
We need to show that f (a, b, c) ≥ 0 and g(a, b, c) ≥ 0. Indeed,
X X X X
f (a, b, c) = ( a2 )2 − 3a bc a≥( a b)2 − 3a bc a≥0
D
A
and
g(a, b, c) =
X
X
PI
[2(a4 + b4 ) + 10a2 b2 − 7a b(a2 + b2 )]
M
= (a − b)2 (2a2 − 3a b + 2b2 ) ≥ 0.
LY
The equality occurs for a = b = c.

Second Solution. Write the inequality in P 1.82 as
O
X 2a2 + (2k + 1)bc

3
2− ≤ ,
.M
b2 + k bc + c 2 k+2
X 2(b2 + c 2 ) − bc − 2a2 3
W
≤ .
b2 + k bc + c2 k+2
Since b + c ≥ 2bc, we get
2 2
W
3bc − 2a2 3
W
X
≤ ,
b2 + k bc + c 2 k+2
which is just the desired inequality.
P 1.84. If a, b, c are nonnegative real numbers, no two of which are zero, then
a2 + 16bc b2 + 16ca c 2 + 16a b

+ 2 + 2 ≥ 10.
b2 + c 2 c + a2 a + b2
Solution. Let a ≤ b ≤ c, and
a2 + 16bc b2 + 16ca c 2 + 16a b

E(a, b, c) = + 2 + 2 .
b2 + c 2 c + a2 a + b2
Consider two cases.
Case 1: 16b3 ≥ ac 2 . We will show that
E(a, b, c) ≥ E(0, b, c) ≥ 10.
We have
L
a2 a(16c 3 − a b2 ) a(16b3 − ac 2 )
E(a, b, c) − E(0, b, c) = + + 2 2 ≥ 0,
.M
b2 + c 2 c 2 (c 2 + a2 ) b (a + b2 )
since c 3 − a b2 ≥ 0 and 16b3 − ac 2 ≥ 0. Also,
D
b2 c 2
A
16bc
E(0, b, c) − 10 = + + 2 − 10
b2 + c 2 c 2 b
=
PI
(b − c)4 (b2 + c 2 + 4bc)
≥ 0.
b2 c 2 (b2 + c 2 )
M
Case 2: 16b3 ≤ ac 2 . It suffices to show that

LY
c 2 + 16a b
O
≥ 10.
a2 + b2
.M
Indeed,
16b3
+ 16a b
W
c + 16a b
2
a
− 10 ≥ − 10
a2 + b2 a2 + b2
W
16b
= − 10 ≥ 16 − 10 > 0.
a
W
This completes the proof. The equality holds for a = 0 and b = c (or any cyclic permu-
tation).
a2 + 128bc b2 + 128ca c 2 + 128a b

+ + ≥ 46.
b2 + c 2 c 2 + a2 a2 + b2
a2 + 128bc b2 + 128ca c 2 + 128a b

E(a, b, c) = + + .
b2 + c 2 c 2 + a2 a2 + b2
Consider two cases.
Case 1: 128b3 ≥ ac 2 . We will show that
E(a, b, c) ≥ E(0, b, c) ≥ 46.
We have
a2 a(128c 3 − a b2 ) a(128b3 − ac 2 )
L
E(a, b, c) − E(0, b, c) = + + ≥ 0,
b2 + c 2 c 2 (c 2 + a2 ) b2 (a2 + b2 )
.M
since c 3 − a b2 ≥ 0 and 128b3 − ac 2 ≥ 0. Also,
D
128bc b2 c 2
A
E(0, b, c) − 46 = + + 2 − 46
b2 + c 2 c 2 b
=
2 PI
(b + c − 4bc) (b + c 2 + 8bc)
2 2 2
≥ 0.
b2 c 2 (b2 + c 2 )
M
Case 2: 128b3 ≤ ac 2 . It suffices to show that
LY
c 2 + 128a b
≥ 46.
O
a2 + b2
Indeed,
.M
128b3
+ 128a b
c + 128a b
2
a
W
− 46 ≥ − 46
a2 + b2 a2 + b2
128b
W
= − 46 ≥ 128 − 46 > 0.
a
W
b c
This completes the proof. The equality holds for a = 0 and + = 4 (or any cyclic
c b
permutation).
a2 + 64bc b2 + 64ca c 2 + 64a b

+ + ≥ 18.
(b + c)2 (c + a)2 (a + b)2
a2 + 64bc b2 + 64ca c 2 + 64a b

E(a, b, c) = + + .
(b + c)2 (c + a)2 (a + b)2
Consider two cases.
Case 1: 64b3 ≥ c 2 (a + 2b). We will show that
E(a, b, c) ≥ E(0, b, c) ≥ 18.
We have
a2 a[64c 3 − b2 (a + 2c)] a[64b3 − c 2 (a + 2b)]
L
E(a, b, c) − E(0, b, c) = + +
(b + c)2 c 2 (c + a)2 b2 (a + b)2
.M
a[64c 3 − b2 (a + 2c]) a[64b2 c − b2 (c + 2c)] 61a b2 c
≥ ≥ = ≥ 0.
D
c 2 (c + a)2 c 2 (c + a)2 c 2 (c + a)2
A
Also,
E(0, b, c) − 18 =
64bc
+
PI
b2 c 2
+ 2 − 18
(b + c)2 c 2 b
M
(b − c) (b + c + 6bc)
4 2 2
= ≥ 0.
LY
b2 c 2 (b + c)2
Case 2: 64b3 ≤ c 2 (a + 2b). It suffices to show that

O
c 2 + 64a b
.M
≥ 18.
(a + b)2
Indeed,
W
64b3
+ 64a b
W
c + 64a b
2
a + 2b
− 18 ≥ − 18
(a + b)2 (a + b)2
W
64b 64
= − 18 ≥ − 18 > 0.
a + 2b 3
This completes the proof. The equality holds for a = 0 and b = c (or any cyclic permu-
tation).
P 1.87. Let a, b, c be nonnegative real numbers, no two of which are zero. If k ≥ −1, then
≥ a + b + c.
b2 + k bc + c 2
Solution. We apply the SOS method. Write the inequality as follows
− a ≥ 0,
b2 + k bc + c 2
X a(a b + ac − b2 − c 2 )
≥ 0,
b2 + k bc + c 2
X a b(a − b) X ac(a − c)
+ ≥ 0,
b2 + k bc + c 2 b2 + k bc + c 2
L
+ ≥ 0,
b2 + k bc + c 2 c 2 + kca + a2
.M
X
a b(a2 + ka b + b2 )(a + b + kc)(a − b)2 ≥ 0.
D
Without loss of generality, assume that a ≥ b ≥ c. Since a + b + kc ≥ a + b − c > 0, it
A
PI
b(b2 + k bc + c 2 )(b + c + ka)(b − c)2 + a(c 2 + kca + a2 )(c + a + k b)(c − a)2 ≥ 0.
M
Since c + a + k b ≥ c + a − b ≥ 0 and c 2 + kca + a2 ≥ b2 + k bc + c 2 , it is enough to prove
LY
that
b(b + c + ka)(b − c)2 + a(c + a + k b)(c − a)2 ≥ 0.
O
We have
.M
b(b + c + ka)(b − c)2 + a(c + a + k b)(c − a)2 ≥
≥ [b(b + c + ka) + a(c + a + k b)](b − c)2

W
= [a2 + b2 + 2ka b + c(a + b)](b − c)2

W
≥ [(a − b)2 + c(a + b)](b − c)2 ≥ 0.

W
−3
P 1.88. Let a, b, c be nonnegative real numbers, no two of which are zero. If k ≥ , then
2
X a3 + (k + 1)a bc
≥ a + b + c.
b2 + k bc + c 2

Solution. Use the SOS method. Write the inequality as follows

X a3 + (k + 1)a bc X a3 − a(b2 − bc + c 2 )
− a ≥ 0, ≥ 0,
b2 + k bc + c 2 b2 + k bc + c 2
X (b + c)a3 − a(b3 + c 3 ) X a b(a2 − b2 ) + ac(a2 − c 2 )

≥ 0, ≥ 0,
(b + c)(b2 + k bc + c 2 ) (b + c)(b2 + k bc + c 2 )
X a b(a2 − b2 ) X ba(b2 − a2 )
+ ≥ 0,
(b + c)(b2 + k bc + c 2 ) (c + a)(c 2 + kca + a2 )
X
(a2 − b2 )2 a b(a2 + ka b + b2 )[a2 + b2 + a b + (k + 1)c(a + b + c)] ≥ 0,
L
.M
X
(b2 − c 2 )2 bc(b2 + k bc + c 2 )Sa ≥ 0,
where
D
Sa = b2 + c 2 + bc + (k + 1)a(a + b + c).
A
Since Sc > 0, it suffices to show that
PI
(b2 − c 2 )2 b(b2 + k bc + c 2 )Sa + (c 2 − a2 )2 a(c 2 + kca + a2 )S b ≥ 0.
M
Since (c 2 − a2 )2 ≥ (b2 − c 2 )2 , a ≥ b,
LY
c 2 + kca + a2 − (b2 + k bc + c 2 ) = (a − b)(a + b + kc) ≥ 0,

O
and
.M
1
S b = a2 + c 2 + ac + (k + 1)b(a + b + c) ≥ a2 + c 2 + ac − b(a + b + c)
2
W
(a − b)(2a + b) + c(2a + 2c − b)
= ≥ 0,
2
W
it is enough to show that Sa + S b ≥ 0. Indeed,

W
Sa + S b = a2 + b2 + 2c 2 + c(a + b) + (k + 1)(a + b)(a + b + c)

1
≥ a2 + b2 + 2c 2 + c(a + b) − (a + b)(a + b + c)
2
(a − b)2 + c(a + b + 4c)
= ≥ 0.
2
This completes the proof. The equality holds for a = b = c, and for a = 0 and b = c (or
P 1.89. Let a, b, c be nonnegative real numbers, no two of which are zero. If k > 0, then
2a k − b k − c k 2b k − c k − a k 2c k − a k − b k
+ 2 + 2 ≥ 0.
b2 − bc + c 2 c − ca + a2 a − a b + b2
Solution. Let
X = bk − c k , Y = c k − ak , Z = ak − bk ,
A = b2 − bc + c 2 , B = c 2 − ca + a2 , C = a2 − a b + b2 .
L
Without loss of generality, assume that a ≥ b ≥ c. This involves A ≤ B, A ≤ C, X ≥ 0,
.M
and Z ≥ 0. Since
X + 2Z X − Z 2X + Z
D
X 2a k − b k − c k
= + −
b2 − bc + c2 A B C
A
1 1 2 2 1 1

=X + − +Z PI
− − ,
A B C A B C
M
1 1 2
LY
+ − ≥ 0.
A B C
O
1 1 1 1
− ≥ − ,
.M
A C C B
that is,
W
(a − c)(a + c − b)(a2 − ac + c 2 ) ≥ (b − c)(a − b − c)(b2 − bc + c 2 ).

W
For the nontrivial case a > b + c, we can obtain this inequality from
W
a + c − b ≥ a − b − c,
a − c ≥ b − c,
a2 − ac + c 2 > b2 − bc + c 2 .
This completes the proof. The equality holds for a = b = c, and for a = 0 and b = c (or
P 1.90. If a, b, c are the lengths of the sides of a triangle, then
b+c−a c+a−b a+b−c 2(a + b + c)

(a) + 2 + 2 ≥ 2 ;
b2− bc + c 2 c − ca + a 2 a − ab + b 2 a + b2 + c 2
a2 − 2bc b2 − 2ca c 2 − 2a b
(b) + + ≤ 0.
b2 − bc + c 2 c 2 − ca + a2 a2 − a b + b2
Solution. (a) By the Cauchy-Schwarz inequality, we get
[ (b + c − a)]2
P
X b+c−a
L
≥P
b2 − bc + c 2 (b + c − a)(b2 − bc + c 2 )
.M
( a)2
P
= P .
2 a3 − a2 (b + c) + 3a bc
P
D
A
On the other hand, from
PI
(b + c − a)(c + a − b)(a + b − c) ≥ 0,
M
we get X X
a2 (b + c) − a3 ,
LY
2a bc ≤
and hence
O
a3 + a2 (b + c) ( a)( a2 )
P P P P
X X
3 2
2 a − a (b + c) + 3a bc ≤ = .
.M
2 2
Therefore, P
b+c−a 2 a
W
X
≥ P .
b2 − bc + c 2 a2
W
The equality holds for a degenerate triangle with a = b + c (or any cyclic permutation).
W
(b) Since
a2 − 2bc
X (b − c)2 + (b + c)2 − a2
= 2 − ,
b2 − bc + c 2 b2 − bc + c 2
X (b − c)2 X b+c−a
+ (a + b + c) ≥ 6.
b2 − bc + c 2 b2 − bc + c 2
Using the inequality in (a), it suffices to prove that
X (b − c)2 2(a + b + c)2
+ ≥ 6.
b2 − bc + c 2 a2 + b2 + c 2

X (b − c)2 X 2(b − c)2
≥ ,
b2 − bc + c 2 a2 + b2 + c 2
or, equivalently,
X (b − c)2 (a − b + c)(a + b − c)
≥ 0,
b2 − bc + c 2
which is true. The equality holds for degenerate triangles with either a/2 = b = c (or
any cyclic permutation), or a = 0 and b = c (or any cyclic permutation).
Remark. The following generalization of the inequality in (b) holds (Vasile Cîrtoaje,
L
2009):
.M
• Let a, b, c be the lengths of the sides of a triangle. If k ≥ −1, then
X a2 − 2(k + 2)bc
D
≤ 0.
b2 + k bc + c 2
A
with equality for a = 0 and b = c (or any cyclic permutation).
PI
M
LY
a2 b2 c2 1
+ + ≤ .
5a + (b + c)
2 2 5b + (c + a)
2 2 5c + (a + b)
2 2
O
3
(Vo Quoc Ba Can, 2009)
.M
Solution. Apply the Cauchy-Schwarz inequality in the following manner

9 (1 + 2)2 1 2
W
= ≤ 2 + 2 .
5a + (b + c)
2 2 (a + b + c ) + 2(2a + bc)
2 2 2 2 a +b +c
2 2 2a + bc
W
Then,
W
X 9a2 X a2 X 2a2 X bc
≤ + = 4 − ,
5a2 + (b + c)2 a2 + b2 + c 2 2a2 + bc 2a2 + bc
and it remains to show that X bc
≥ 1.
+ bc
2a2
This is a known inequality, which can be proved by the Cauchy-Schwarz inequality, as
follows
( bc)2
P
X bc
≥P = 1.
2a2 + bc bc(2a2 + bc)
b2 + c 2 − a2 c 2 + a2 − b2 a2 + b2 − c 2 1
+ + ≥ .
2a + (b + c)
2 2 2b + (c + a)
2 2 2c + (a + b)
2 2 2

X b2 + c 2 − a2 1

− ≥ 0,
2a2 + (b + c)2 6
L
X 5(b2 + c 2 − 2a2 ) + 2(a2 − bc)
≥ 0;
.M
2a2 + (b + c)2
X 5(b2 − a2 ) + 5(c 2 − a2 ) + (a − b)(a + c) + (a − c)(a + b)
≥ 0;
D
2a2 + (b + c)2
A
X (b − a)[5(b + a) − (a + c)] X (c − a)[5(c + a) − (a + b)]
+ ≥ 0;
2a2 + (b + c)2 PI
2a2 + (b + c)2
X (b − a)[5(b + a) − (a + c)] X (a − b)[5(a + b) − (b + c)]
+
M
≥ 0;
2a2 + (b + c)2 2b2 + (c + a)2
LY
X
(a − b)2 [2c 2 + (a + b)2 ][2(a2 + b2 ) + c 2 + 3a b − 3c(a + b)] ≥ 0,
X
O
(b − c)2 R a Sa ≥ 0,
where
.M
R a = 2a2 + (b + c)2 , Sa = a2 + 2(b2 + c 2 ) + 3bc − 3a(b + c).

Without loss of generality, assume that a ≥ b ≥ c. We have
W
S b = b2 + 2(c 2 + a2 ) + 3ca − 3b(c + a) = (a − b)(2a − b) + 2c 2 + 3c(a − b) ≥ 0,

W
Sc = c 2 + 2(a2 + b2 ) + 3a b − 3c(a + b) ≥ 7a b − 3c(a + b) ≥ 3a(b − c) + 3b(a − c) ≥ 0,

W
Sa + S b = 3(a − b)2 + 4c 2 ≥ 0.
Since X
(b − c)2 R a Sa ≥ (b − c)2 R a Sa + (c − a)2 R b S b
= (b − c)2 R a (Sa + S b ) + [(c − a)2 R b − (b − c)2 R a ]S b ,

(a − c)2 R b ≥ (b − c)2 R a .
We can get this by multiplying the inequalities
b2 (a − c)2 ≥ a2 (b − c)2
and
a2 R b ≥ b2 R a .
P 1.93. Let a, b, c be positive real numbers. If k > 0, then
3a2 − 2bc 3b2 − 2ca 3c 2 − 2a b 3

+ + ≤ .
ka + (b − c)
2 2 k b + (c − a)
2 2 kc + (a − b)
2 2 k
L
.M
D
X1 3a2 − 2bc

− ≥ 0,
A
k ka2 + (b − c)2
PI
X b2 + c 2 − 2a2 + 2(k − 1)(bc − a2 )
≥ 0;
M
ka2 + (b − c)2
X (b2 − a2 ) + (c 2 − a2 ) + (k − 1)[(a + b)(c − a) + (a + c)(b − a)]
LY
≥ 0;
ka2 + (b − c)2
O
X (b − a)[b + a + (k − 1)(a + c)] X (c − a)[c + a + (k − 1)(a + b)]

+ ≥ 0;
ka2 + (b − c)2 ka2 + (b − c)2
.M
X (b − a)[b + a + (k − 1)(a + c)] X (a − b)[a + b + (k − 1)(b + c)]

+ ≥ 0;
ka2 + (b − c)2 k b2 + (c − a)2
W
X
(a − b)2 [kc 2 + (a − b)2 ][(k − 1)c 2 + 2c(a + b) + (k2 − 1)(a b + bc + ca)] ≥ 0.
W
For k ≥ 1, the inequality is clearly true. Consider further that 0 < k < 1. Since
W
(k − 1)c 2 + 2c(a + b) + (k2 − 1)(a b + bc + ca) >
> −c 2 + 2c(a + b) − (a b + bc + ca) = (b − c)(c − a),

X
(a − b)(b − c)(c − a) (a − b)[kc 2 + (a − b)2 ] ≥ 0.
Since
X X X
(a − b)[kc 2 + (a − b)2 ] = k (a − b)c 2 + (a − b)3
= (3 − k)(a − b)(b − c)(c − a),
we have X
(a − b)(b − c)(c − a) (a − b)[kc 2 + (a − b)2 ] =
= (3 − k)(a − b)2 (b − c)2 (c − a)2 ≥ 0.

This completes the proof. The equality holds for a = b = c.
p
P 1.94. Let a, b, c be nonnegative real numbers, no two of which are zero. If k ≥ 3 + 7,
then
L
a b c 9
(a) + 2 + 2 ≥ ;
.M
a2 + k bc b + kca c + ka b (1 + k)(a + b + c)
1 1 1 9
D
(b) + 2 + 2 ≥ .
ka2 + bc k b + ca kc + a b (k + 1)(a b + bc + ca)
A
PI
Solution. (a) Assume that a = max{a, b, c}. Setting t = (b + c)/2, t ≤ a, by the
M
Cauchy-Schwarz inequality, we get
LY
b c (b + c)2 4t 2
+ ≥ =
b2 + kca c 2 + ka b b(b2 + kca) + c(c 2 + ka b) 8t 3 − 6bc t + 2ka bc
O
2t 2 2t 2 2
= 3 ≥ 3 = .
4t + (ka − 3t)bc 4t + (ka − 3t)t t + ka
.M
On the other hand,

a a
W
≥ 2 .
a2 + k bc a + kt 2
W
a 2 9
W
+ ≥ ,
a2 + kt 2 t + ka (k + 1)(a + 2t)
(a − t)2 [(k2 − 6k + 2)a + k(4k − 5)t] ≥ 0.
This inequality is true, since k2 − 6k + 2 ≥ 0 and 4k − 5 > 0. The equality holds for
a = b = c.
(b) For a = 0, the inequality becomes
1 1 k(8 − k)
+ 2≥ .
b 2 c (k + 1)bc
We have
1 1 k(8 − k) 2 k(8 − k) k2 − 6k + 2
+ − ≥ − = ≥ 0.
b2 c 2 (k + 1)bc bc (k + 1)bc (k + 1)bc
For a, b, c > 0, the desired inequality follows from the inequality in (a) by substituting
p 1/a, 1/b, 1/c, respectively. The equality holds for a = b = c. In the case
a, b, c with
k = 3 + 7, the equality also holds for a = 0 and b = c (or any cyclic permutation).
L
1 1 1 6
+ + ≥ 2 .
.M
2a2 + bc 2b2 + ca 2c 2 + a b a + b2 + c 2 + a b + bc + ca
D
A
1 (b + c)2 4(a + b + c)2
X
2a2 + bc
=
X
(b + c)2 (2a2 + bc)
≥ P PI
(b + c)2 (2a2 + bc)
.
M
LY
X
2(a + b + c)2 (a2 + b2 + c 2 + a b + bc + ca) ≥ 3 (b + c)2 (2a2 + bc),
O
X X X X
.M
2 a4 + 3 a b(a2 + b2 ) + 2a bc a ≥ 10 a2 b2 .
This follows by adding Schur’s inequality

W
X X X
2 a4 + 2a bc a≥2 a b(a2 + b2 )
W
to the inequality X X
W
5 a b(a2 + b2 ) ≥ 10 a2 b2 .
tion).
1 1 1 1
+ + ≥ .
22a2 + 5bc 22b + 5ca 22c + 5a b
2 2 (a + b + c)2

X 1 X (b + c)2 4(a + b + c)2
= ≥ .
22a2 + 5bc (b + c)2 (22a2 + 5bc) (b + c)2 (22a2 + 5bc)
P

X
4(a + b + c)4 ≥ (b + c)2 (22a2 + 5bc),
X X X X
4 a4 + 11 a b(a2 + b2 ) + 4a bc a ≥ 30 a2 b2 .
L
.M
X X X
4 a4 + 4a bc a≥4 a b(a2 + b2 )
D
to the inequality
A
X X
15 a b(a2 + b2 ) ≥ 30 a2 b2 .
PI
M
LY
O
1 1 1 8
+ + ≥ .
.M
2a2 + bc 2b2 + ca 2c 2 + a b (a + b + c)2

W
First Solution. Applying the Cauchy-Schwarz inequality, we have

W
X 1 X (b + c)2 4(a + b + c)2

= ≥ .
W
2a2 + bc (b + c)2 (2a2 + bc) (b + c)2 (2a2 + bc)

P

X
(a + b + c)4 ≥ 2 (b + c)2 (2a2 + bc),
X X X X
a4 + 2 ab(a2 + b2 ) + 4a bc a≥6 a2 b2 .
We will prove the sharper inequality

X X X X
a4 + 2 a b(a2 + b2 ) + a bc a≥6 a2 b2 .

X X X
a4 + a bc a≥ a b(a2 + b2 )
to the inequality X X
3 a b(a2 + b2 ) ≥ 6 a2 b2 .
Second Solution. Without loss of generality, we may assume that a ≥ b ≥ c. Since the
equality holds for c = 0 and a = b, write the inequality as
1 1 1 1 8
L
+ 2 + 2 + 2 ≥
2a2 + bc 2b + ca 4c + 2a b 4c + 2a b (a + b + c)2
.M
and then apply the Cauchy-Schwarz inequality. It suffices to prove that
D
16 8
≥ ,
A
(2a2 + bc) + (2b2 + ca) + (4c + 2a b) + (4c + 2a b) (a + b + c)2
2 2
which is equivalent to the obvious inequality PI

M
c(a + b − 2c) ≥ 0.
LY
O
.M
1 1 1 12
+ + ≥ .
a2 + bc b2 + ca c 2 + a b (a + b + c)2
W

W
Solution. Due to homogeneity, we may assume that a + b + c = 1. On this assumption,

W
write the inequality as

X 1
− 1 ≥ 9,
a2 + bc
X 1 − a2 − bc
≥ 9.
a2 + bc
Since
1 − a2 − bc = (a + b + c)2 − a2 − bc > 0,
by the Cauchy-Schwarz inequality, we have
[ (1 − a2 − bc)]2
P
X 1 − a2 − bc
≥P .
a2 + bc (1 − a2 − bc)(a2 + bc)
Then, it suffices to prove that

X X X X
(3 − a2 − bc)2 ≥ 9 (a2 + bc) − 9 (a2 + bc)2 ,
(1 − 4q)(4 − 7q) + 36a bc ≥ 0,
where q = a b + bc + ca. For q ≤ 1/4, this inequality is clearly true. Consider further
that q > 1/4. By Schur’s inequality of degree three
L
we get 1 + 9a bc ≥ 4q, and hence 36a bc ≥ 16q − 4. Thus,
.M
(1 − 4q)(4 − 7q) + 36a bc ≥ (1 − 4q)(4 − 7q) + 16q − 4 = 7q(4q − 1) > 0.
D
A
PI
M
LY
1 1 1 1 2
(a) + 2 + 2 ≥ 2 + ;
a2 + 2bc b + 2ca c + 2a b a +b +c
2 2 a b + bc + ca
O
a(b + c) b(c + a) c(a + b) a b + bc + ca

.M
(b) + 2 + 2 ≥1+ 2 .
a + 2bc b + 2ca c + 2a b
2 a + b2 + c 2
(Darij Grinberg and Vasile Cîrtoaje, 2005)

W
Solution. (a) Write the inequality as

W
(b + 2ca)(c 2 + 2a b)
P 2
a b + bc + ca + 2a2 + 2b2 + 2c 2
W
≥ .
(a2 + 2bc)(b2 + 2ca)(c 2 + 2a b) (a2 + b2 + c 2 )(a b + bc + ca)
Since
X
(b2 + 2ca)(c 2 + 2a b) = (a b + bc + ca)(a b + bc + ca + 2a2 + 2b2 + 2c 2 ),
(a2 + b2 + c 2 )(a b + bc + ca)2 ≥ (a2 + 2bc)(b2 + 2ca)(c 2 + 2a b),
which is just the inequality (a) in P 2.16 in Volume 1. The equality holds for a = b, or
b = c, or c = a.
(b) Write the inequality in (a) as

X a b + bc + ca a b + bc + ca
≥2+ ,
a2 + 2bc a2 + b2 + c 2
or X a(b + c) X bc a b + bc + ca
+ ≥2+ 2 .
a + 2bc
2 a + 2bc
2 a + b2 + c 2
The desired inequality follows by adding this inequality to
X bc
1≥ .
a2 + 2bc
L
The last inequality is equivalent to
.M
X a2
≥ 1,
a2 + 2bc
D
which follows by applying the AM-GM inequality as follows
A
X a2
a2 + 2bc
≥
X a2
a2 + b2 + c 2
PI
= 1.
M
LY
O
.M
a b c a+b+c
(a) + 2 + 2 ≤ ;
W
a(b + c) b(c + a) c(a + b) a2 + b2 + c 2

+ + +
W
(b) ≤ 1 .
W
Solution. (a) Use the SOS method. Write the inequality as

X a b + bc + ca

a 1− ≥ 0,
a2 + 2bc
X a(a − b)(a − c)
≥ 0.
a2 + 2bc
Assume that a ≥ b ≥ c. Since (c − a)(c − b) ≥ 0, it suffices to show that
a(a − b)(a − c) b(b − a)(b − c)

+ ≥ 0.
a2 + 2bc b2 + 2ca
c(a − b)2 [2a(a − c) + 2b(b − c) + 3a b] ≥ 0,
which is clearly true. The equality holds for a = b = c, and for a = 0 and b = c (or any
cyclic permutation).
(b) Since
a(b + c) a(a + b + c) a2
= − ,
a2 + 2bc a2 + 2bc a2 + 2bc
L
X a a2 + b2 + c 2 X a2
(a + b + c) + +
.M
≤ 1 .
a2 + 2bc a b + bc + ca a2 + 2bc
According to the inequality in (a), it suffices to show that
D
(a + b + c)2 a2 + b2 + c 2 X a2
A
≤1+ + ,
a b + bc + ca a b + bc + ca a2 + 2bc
PI
M
X a2
≥ 1.
a2 + 2bc
LY
Indeed,
a2 a2
O
X X
≥ = 1.
a2 + 2bc a2 + b2 + c 2
.M

W
W
a b c a+b+c
W
(a) + 2 + 2 ≥ 2 ;
2a2 + bc 2b + ca 2c + a b a + b2 + c 2
b+c c+a a+b 6
(b) + 2 + 2 ≥ .
2a + bc 2b + ca 2c + a b
2 a+b+c
Solution. Assume that a ≥ b ≥ c.

(a) Multiplying by a + b + c, we can write the inequality as follows
X a(a + b + c) (a + b + c)2
≥ ,
2a2 + bc a2 + b2 + c 2
(a + b + c)2 X a(a + b + c)

3− 2 ≥ 1− ,
a + b2 + c 2 2a2 + bc
X X (a − b)(a − c)
2 (a − b)(a − c) ≥ (a2 + b2 + c 2 ) ,
2a2 + bc
X 3a2 − (b − c)2
(a − b)(a − c) ≥ 0,
2a2 + bc
3 f (a, b, c) + (a − b)(b − c)(c − a)g(a, b, c) ≥ 0,
where
X a2 (a − b)(a − c) b−c X
f (a, b, c) = , g(a, b, c) = .
+ bc
2a2 2a2 + bc
L
It suffices to show that f (a, b, c) ≥ 0 and g(a, b, c) ≤ 0. We have
.M
a2 (a − b)(a − c) b2 (b − a)(b − c)
f (a, b, c) ≥ +
D
2a2 + bc 2b2 + ca
a (a − b)(b − c) b (b − a)(b − c)
2 2
A
≥ +
2a2 + bc 2b2 + ca
=
2 2 PI
a c(a − b) (b − c)(a + a b + b2 )
2
≥ 0.
(2a2 + bc)(2b2 + ca)
M
Also,
LY
b−c (a − b) + (b − c) a−b
g(a, b, c) = − + 2
O
2a + bc
2 2b + ca
2
2c + a b
1 1 1 1

= (a − b) − + (b − c) −
.M
2c 2 + a b 2b2 + ca 2a2 + bc 2b2 + ca

(a − b)(b − c) 2b + 2c − a 2b + 2a − c

= − =
2b2 + ca 2c 2 + a b 2a2 + bc
W
2(a − b)(b − c)(c − a)(a2 + b2 + c 2 − a b − bc − ca)

= ≤ 0.
W
(2a2 + bc)(2b2 + ca)(2c 2 + a b)

W
(b) We apply the SOS method. Write the inequality as follows

X (b + c)(a + b + c)
− 2 ≥ 0,
2a2 + bc
X (b2 + a b − 2a2 ) + (c 2 + ca − 2a2 )
≥ 0,
2a2 + bc
X (b − a)(b + 2a) + (c − a)(c + 2a)
≥ 0,
2a2 + bc
X (b − a)(b + 2a) X (a − b)(a + 2b)
+ ≥ 0,
2a2 + bc 2b2 + ca
a + 2b b + 2a
X
(a − b) − ≥ 0,
2b2 + ca 2a2 + bc
X
(a − b)2 (2c 2 + a b)(a2 + b2 + 3a b − ac − bc) ≥ 0.
Since
a2 + b2 + 3a b − ac − bc ≥ a2 + b2 + 2a b − ac − bc = (a + b)(a + b − c),

X
(a − b)2 (2c 2 + a b)(a + b)(a + b − c) ≥ 0.
L
.M
This inequality is true if
D
(b − c)2 (2a2 + bc)(b + c)(b + c − a) + (c − a)2 (2b2 + ca)(c + a)(c + a − b) ≥ 0;
A
that is,
PI
(a − c)2 (2b2 + ca)(a + c)(a + c − b) ≥ (b − c)2 (2a2 + bc)(b + c)(a − b − c).
M
LY
Since a + c ≥ b + c and a + c − b ≥ a − b − c, it is enough to prove that
(a − c)2 (2b2 + ca) ≥ (b − c)2 (2a2 + bc).

O
.M
We can obtain this inequality by multiplying the inequalities
b2 (a − c)2 ≥ a2 (b − c)2
W
and
W
a2 (2b2 + ca) ≥ b2 (2a2 + bc).

W
a(b + c) b(c + a) c(a + b) (a + b + c)2

+ + ≥ .
a2 + bc b2 + ca c2 + a b a2 + b2 + c 2
(Pham Huu Duc, 2006)

Solution. Assume that a ≥ b ≥ c and write the inequality as follows
(a + b + c)2 X a b + ac

3− ≥ 1 − ,
a2 + b2 + c 2 a2 + bc
X X (a − b)(a − c)
2 (a − b)(a − c) ≥ (a2 + b2 + c 2 ) ,
a2 + bc
X (a − b)(a − c)(a + b − c)(a − b + c)
≥ 0.
a2 + bc
(b − c)(b − a)(b + c − a)(b − c + a) (c − a)(c − b)(c + a − b)(c − a + b)
L
+ ≥ 0,
b2 + ca c2 + a b
.M
D
(b − c)2 (c − a + b)2 (a2 + bc)
≥ 0.
A
(b2 + ca)(c 2 + a b)
PI
M
LY
P 1.103. Let a, b, c be nonnegative real numbers, no two of which are zero. If k > 0, then
O
p p p p
b2 + c 2 + 3bc c 2 + a2 + 3ca a2 + b2 + 3a b 3(2 + 3)
+ + ≥ .
a2 + k bc b2 + kca c 2 + ka b 1+k
.M

W
Solution. Write the inequality in the form f6 (a, b, c) ≥ 0, where

p
W
X
f6 (a, b, c) = (1 + k) (b2 + c 2 + 3bc)(b2 + kca)(c 2 + ka b)
W
p
−3(2 + 3 )(a2 + k bc)(b2 + kca)(c 2 + ka b).
Clearly, f6 (a, b, c) has the same highest coefficient as f (a, b, c), where
Xp
f (a, b, c) = (1 + k) ( 3bc − a2 )(b2 + kca)(c 2 + ka b)
p
−3(2 + 3 )(a2 + k bc)(b2 + kca)(c 2 + ka b));
therefore,
p p
A = 3(1 + k)3 ( 3 − 1) − 3(2 + 3 )(1 + k)3
= −9(1 + k)3 .
Since A ≤ 0, according to P 3.76-(a) in Volume 1, it suffices to prove the original in-

equality for b = c = 1 and for a = 0.
In the first case, this inequality is equivalent to
p p 2
+

3 1 3
(a − 1)2 (k + 1)a2 − 1 + (k − 2)a + k − ≥ 0.
2 2
For the nontrivial case k > 2, we have

p 2 p
1+ 3 1+ 3

p
2
(k + 1)a + k − ≥2 k+1 k− a
2 2
L
p p
.M
p 1+ 3

3
≥2 3 k− a ≥ 1+ (k − 2)a.
2 2
D
In the second case (a = 0), the original inequality can be written as
A
p
3(2 + 3)
2
1 b c p c2

b PI
+ + 3 + 2 + 2 ≥ ,
k c b c b 1+k
M
and is true if p p
2+ 3 3(2 + 3)
LY
+2≥ ,
k 1+k
O
p 2
1+ 3

.M
k− ≥ 0.
2
p
1+ 3
W
The equality holds for a = b = c. If k = , then the equality holds also for a = 0
2
W
W
1 1 1 8 6
+ 2 + 2 + 2 ≥ .
a2 +b 2 b +c 2 c +a 2 a +b +c
2 2 a b + bc + ca
Solution. Multiplying by a2 + b2 + c 2 , the inequality becomes
a2 b2 c2 6(a2 + b2 + c 2 )
+ + + 11 ≥ .
b2 + c 2 c 2 + a2 a2 + b2 a b + bc + ca
Since
a2 b2 c2

+ + (a2 b2 + b2 c 2 + c 2 a2 ) =
b2 + c 2 c 2 + a2 a2 + b2
1 1 1

4 4 4 2 2 2
=a +b +c +a b c + + ≥ a4 + b4 + c 4 ,
a2 + b2 b2 + c 2 c 2 + a2
a4 + b4 + c 4 6(a2 + b2 + c 2 )
+ 11 ≥ ,
a2 b2 + b2 c 2 + c 2 a2 a b + bc + ca
L
(a2 + b2 + c 2 )2 6(a2 + b2 + c 2 )
.M
+ 9 ≥ .
a2 b2 + b2 c 2 + c 2 a2 a b + bc + ca
D
Clearly, it is enough to prove that
A
2
a2 + b2 + c 2 6(a2 + b2 + c 2 )

+9≥ PI ,
a b + bc + ca a b + bc + ca
M
which is 2
a2 + b2 + c 2

LY
−3 ≥ 0.
a b + bc + ca
b c
O
The equality holds for a = 0 and + = 3 (or any cyclic permutation).

c b
.M
W
a(b + c) b(c + a) c(a + b)

W
+ 2 + 2 ≤ 2.
a + 2bc b + 2ca c + 2a b
2
W
(Vo Quoc Ba Can and Vasile Cîrtoaje, 2010)

X a b + ac)

1− 2 ≥ 1,
a + 2bc
X a2 + 2bc − a b − ac
≥ 1.
a2 + 2bc
Since
a2 + 2bc − a b − ac = bc − (a − c)(b − a) ≥ |a − c||b − a| − (a − c)(b − a) ≥ 0,

by the Cauchy-Schwarz inequality, we have
[ (a + 2bc − a b − ac)]2
P 2
X a2 + 2bc − a b − ac
≥P .
a2 + 2bc (a2 + 2bc)(a2 + 2bc − a b − ac)

X
(a2 + b2 + c 2 )2 ≥ (a2 + 2bc)(a2 + 2bc − a b − ac),
which reduces to the obvious inequality
a b(a − b)2 + bc(b − c)2 + ca(c − a)2 ≥ 0.
L
.M
The equality holds for an equilateral triangle, and for a degenerate triangle with a = 0
D
A
P 1.106. If a, b, c are real numbers, then PI
a2 − bc b2 − ca c2 − a b
M
+ + ≥ 0.
2a2 + b2 + c 2 2b2 + c 2 + a2 2c 2 + a2 + b2
LY
(Nguyen Anh Tuan, 2005)

O
First Solution. Rewrite the inequality as

.M
X1 a2 − bc

3
− ≤ ,
2 2a2 + b2 + c 2 2
W
X (b + c)2
≤ 3.
2a2 + b2 + c 2
W
If two of a, b, c are zero, then the inequality is trivial. Otherwise, applying the Cauchy-
W
Schwarz inequality, we get

X(b + c)2 X (b + c)2
=
2a2 + b2 + c 2 (a2 + b2 ) + (a2 + c 2 )
X b2 c2
X
b2 X a2
≤ + = + = 3.
a2 + b2 a2 + c 2 a2 + b2 b2 + a2
Second Solution. Use the SOS method. We have
X a2 − bc X (a − b)(a + c) + (a − c)(a + b)
2 =
2a2 + b2 + c 2 2a2 + b2 + c 2
X (a − b)(a + c) X (b − a)(b + c)
= +
2a2 + b2 + c 2 2b2 + c 2 + a2
a+c b+c
X
= (a − b) −
2a2 + b2 + c 2 2b2 + c 2 + a2
X (a − b)2
= (a2 + b2 + c 2 − a b − bc − ca) ≥ 0.
(2a2 + b2 + c 2 )(2b2 + c 2 + a2 )
L
.M
3a2 − bc 3b2 − ca 3c 2 − a b 3
+ + ≤ .
2a + b + c
2 2 2 2b + c + a
2 2 2 2c + a + b
2 2 2 2
D
A
X3 3a2 − bc
PI
− ≥ 3,
M
2 2a2 + b2 + c 2
LY
X 8bc + 3(b − c)2

≥ 6.
2a2 + b2 + c 2
O

.M
3[4bc + (b − c)2 ]2 2(b + c)4

8bc + 3(b − c)2 ≥ = .
6bc + (b − c)2 b2 + c 2 + 4bc
W

W
X (b + c)4
≥ 2.
(2a2 + b2 + c 2 )(b2 + c 2 + 4bc)
W
Using again the Cauchy-Schwarz inequality, we get
[ (b + c)2 ]2
P
X (b + c)4
≥P = 2.
(2a2 + b2 + c 2 )(b2 + c 2 + 4bc) (2a2 + b2 + c 2 )(b2 + c 2 + 4bc)
The equality holds for a = b = c, for a = 0 and b = c, and for b = c = 0 (or any cyclic
permutation).
X1 3a2 − bc

− ≥ 0,
2 2a2 + b2 + c 2
X (b + c + 2a)(b + c − 2a)
≥ 0,
2a2 + b2 + c 2
X (b + c + 2a)(b − a) + (b + c + 2a)(c − a)
≥ 0,
2a2 + b2 + c 2
X (b + c + 2a)(b − a) X (c + a + 2b)(a − b)
+ ≥ 0,
2a2 + b2 + c 2 2b2 + c 2 + a2
c + a + 2b b + c + 2a
X
(a − b) − ≥ 0,
2b2 + c 2 + a2 2a2 + b2 + c 2
X
(3a b + bc + ca − c 2 )(2c 2 + a2 + b2 )(a − b)2 ≥ 0.
L
Since 3a b + bc + ca − c 2 ≥ c(a + b − c), it suffices to show that
.M
X
c(a + b − c)(2c 2 + a2 + b2 )(a − b)2 ≥ 0.
D
Assume that a ≥ b ≥ c. It is enough to prove that
A
a(b + c − a)(2a2 + b2 + c 2 )(b − c)2 + b(c + a − b)(2b2 + c 2 + a2 )(c − a)2 ≥ 0;
that is,
PI
M
b(c + a − b)(2b2 + c 2 + a2 )(a − c)2 ≥ a(a − b − c)(2a2 + b2 + c 2 )(b − c)2 .
LY
Since c + a − b ≥ a − b − c, it suffices to prove that

O
b(2b2 + c 2 + a2 )(a − c)2 ≥ a(2a2 + b2 + c 2 )(b − c)2 .

.M
We can obtain this inequality by multiplying the inequalities
b2 (a − c)2 ≥ a2 (b − c)2
W
and
W
a(2b2 + c 2 + a2 ) ≥ b(2a2 + b2 + c 2 ).
The last inequality is equivalent to
W
(a − b)[(a − b)2 + a b + c 2 ] ≥ 0.
(b + c)2 (c + a)2 (a + b)2

+ + ≥ 2.
4a2 + b2 + c 2 4b2 + c 2 + a2 4c 2 + a2 + b2
[ (b + c)2 ]2
P
X (b + c)2
≥P
4a2 + b2 + c 2 (b + c)2 (4a2 + b2 + c 2 )
2[ a4 + 3 a2 b2 + 4a bc a + 2 a b(a2 + b2 )]
P P P P
= P ≥ 2,
a4 + 5 a2 b2 + 4a bc a + a b(a2 + b2 )
P P P
since X X
a b(a2 + b2 ) ≥ 2 a2 b2 .
The equality holds for a = b = c, and for b = c = 0 (or any cyclic permutation).
L
.M
D
A
P 1 3
(a) ≤ ;
11a2 + 2b + 2c
2 2 PI
5(a b + bc + ca)
1 1 1
M
+
P
(b) ≤ .
4a2 + b2 + c 2 2(a2 + b2 + c 2 ) a b + bc + ca
LY

O
Solution. We will prove that

.M
X k+2 11 − 2k 2(k − 1)
≤ 2 +
ka2 +b +c
2 2 a +b +c
2 2 a b + bc + ca
W
for any k > 1. Due to homogeneity, we may assume that a2 + b2 + c 2 = 3. On this

hypothesis, we need to show that
W
X k+2 11 − 2k 2(k − 1)
≤ + .
W
(k − 1)a + 3
2 3 a b + bc + ca
Using the substitution m = 3/(k − 1), m > 0, the inequality can be written as
X 1 6
m(m + 1) ≤ 3m − 2 + .
a2 +m a b + bc + ca
p p
(a2 + m)[m + (m + 1 − a)2 ] ≥ [a m + m(m + 1 − a)]2 = m(m + 1)2 ,
and hence
m(m + 1) a2 − 1
≤ + m + 2 − 2a,
a2 + m m+1
X 1 X
m(m + 1) ≤ 3(m + 2) − 2 a.
a2 + m
X 6
3(m + 2) − 2 a ≤ 3m − 2 + ;
a b + bc + ca
that is,
(4 − a − b − c)(a b + bc + ca) ≤ 3.
Let p = a + b + c. Since
2(a b + bc + ca) = (a + b + c)2 − (a2 + b2 + c 2 ) = p2 − 3,
L
we get
.M
6 − 2(4 − a − b − c)(a b + bc + ca) = 6 − (4 − p)(p2 − 3)
= (p − 3)2 (p + 2) ≥ 0.
D
This completes the proof. The equality holds for a = b = c.
A
PI
M
P 1.110. If a, b, c are nonnegative real numbers such that a b + bc + ca = 3, then
p p p
LY
a b c 3
+ + ≥ .
b+c c+a a+b 2
O

.M

X pa
P 3/4 2
a 1 X 3/4 2
≥P = a .
W
b+c a(b + c) 6
W

a3/4 + b3/4 + c 3/4 ≥ 3,
W
which follows immediately by Remark 1 from the proof of the inequality in P 3.33 in
Volume 1. The equality occurs for a = b = c = 1.
Remark. Analogously, according to Remark 2 from the proof of P 3.33 in Volume 1, we
can prove that
ak bk ck 3
+ + ≥
b+c c+a a+b 2
4 ln 2 4 ln 2
for all k ≥ 3 − ≈ 0.476. For k = 3 − , the equality occurs for a = b = c = 1,
ln 3 p ln 3
and also for a = 0 and b = c = 3 (or any cyclic permutation).
P 1.111. If a, b, c are nonnegative real numbers such that a b + bc + ca ≥ 3, then
1 1 1 1 1 1
+ + ≥ + + .
2+a 2+ b 2+c 1+ b+c 1+c+a 1+a+ b
Solution. Denote
1 1 1 1 1 1
E(a, b, c) = + + − − − .
2+a 2+ b 2+c 1+ b+c 1+c+a 1+a+ b
Consider first the case a b + bc + ca = 3. We will show that
L
1 1 1 1 1 1
.M
+ + ≥1≥ + + .
2+a 2+ b 2+c 1+ b+c 1+c+a 1+a+ b
D
By direct calculation, we can show that the left inequality is equivalent to a bc ≤ 1.
Indeed, applying the AM-GM inequality, we get
A
p
PI
3 = a b + bc + ca ≥ 3 a b · bc · ca.
M
Also, the right inequality is equivalent to
LY
a + b + c ≥ 2 + a bc.
O
Since a bc ≤ 1, it suffices to show that
a + b + c ≥ 3.
.M
Indeed,
W
(a + b + c)2 ≥ 3(a b + bc + ca) = 9.

Consider further that a b+ bc+ca > 3. Without loss of generality, assume that a ≥ b ≥ c,
W
a > 1. For c ≥ 1, that is, a ≥ b ≥ c ≥ 1, the desired inequality follows by summing the
W
obvious inequalities
1 1
≥ ,
2+a 1+c+a
1 1
≥ ,
2+ b 1+a+ b
1 1
≥ .
2+c 1+ b+c
Therefore, assume now that c < 1. Consider the cases b + c ≥ 2 and b + c < 2.
Case 1: b + c ≥ 2, a > 1, c < 1. We will show that
E(a, b, c) ≥ E(1, b, c) ≥ 0.
We have
1 1 1 1 1 1

E(a, b, c) − E(1, b, c) = − + − + −
2+a 3 2+ b 1+a+ b 2+c 1+c+a
−1 1 1

= (a − 1) + +
3(2 + a) (2 + b)(1 + a + b) (2 + c)(1 + c + a)
−1 1

> (a − 1) +
3(2 + a) (2 + c)(1 + c + a)
(a − 1)(1 − c)(4 + c + a)
= >0
3(2 + a)(2 + c)(1 + c + a)
L
and
b+c−2
.M
E(1, b, c) = ≥ 0.
3(1 + b + c)
D
Case 2: b + c < 2, a > 1, c < 1. From b + c < 2, it follows that
A
2
bc ≤

b+c
2
PI
< 1.
M
For fixed b and c, define the function
LY
f (x) = E(x, b, c).

O
Since
.M
−1 1 1 −1 1
f 0 (x) = + + > +
(2 + x) 2 (1 + c + x)2 (1 + x + b)2 (2 + x)2 (1 + c + x)2
(1 − c)(3 + 2x + c)
W
= > 0,
(2 + x)2 (1 + c + x)2
W
f (x) is strictly increasing for x ≥ 0. Since

W
3 − bc
a> ,
b+c
3 − bc 3 − bc

we have f (a) > f . Therefore, it suffices to prove that f ≥ 0, which
b+c b+c
3 − bc
is equivalent to E(a, b, c) ≥ 0 for a = , that is, for a b + bc + ca = 3. But this was
b+c
proved in the first part of the proof. So, the proof is completed. The equality occurs for
a = b = c = 1.
a2 − bc b2 − ca c2 − a b
(a) + + ≤ 0;
3a2 + b2 + c 2 3b2 + c 2 + a2 3c 2 + a2 + b2
a4 − b2 c 2 b4 − c 2 a2 c 4 − a2 b2
(b) + + ≤ 0.
3a4 + b4 + c 4 3b4 + c 4 + a4 3c 4 + a4 + b4
(Nguyen Anh Tuan and Vasile Cîrtoaje, 2006)
Solution. (a) Apply the SOS method. We have
L
a2 − bc X (a − b)(a + c) + (a − c)(a + b)
.M
X
2 =
3a2 + b2 + c 2 3a2 + b2 + c 2
X (a − b)(a + c) X (b − a)(b + c)
D
= +
3a2 + b2 + c 2 3b2 + c 2 + a2
A
a+c b+c

=
X
(a − b) − PI
3a2 + b2 + c 2 3b2 + c 2 + a2
M
X (a − b)2
= (a2 + b2 + c 2 − 2a b − 2bc − 2ca) .
(3a2 + b2 + c 2 )(3b2 + c 2 + a2 )
LY
Since
O
a2 + b2 + c 2 − 2a b − 2bc − 2ca = a(a − b − c) + b(b − c − a) + c(c − a − b) ≤ 0,

.M
the conclusion follows. The equality holds for an equilateral triangle, and for a degen-
erate triangle with a = 0 and b = c (or any cyclic permutation).
W
(b) Using the same way as above, we get

W
X a4 − b2 c 2 X (a2 − b2 )2
2 = A ,
W
3a4 + b4 + c 4 (3a4 + b4 + c 4 )(3b4 + c 4 + a4 )
where
A = a4 + b4 + c 4 − 2a2 b2 − 2b2 c 2 − 2c 2 a2
= −(a + b + c)(a + b − c)(b + c − a)(c + a − b) ≤ 0.
The equality holds for an equilateral triangle, and for a degenerate triangle with a = b+c

bc ca ab 1
+ 2 + 2 ≥ .
4a2 +b +c
2 2 4b + c + a
2 2 4c + a + b
2 2 2
(Vasile Cîrtoaje and Vo Quoc Ba Can, 2010)

X b2 c 2

2bc X
− ≥ 0,
4a2 + b2 + c 2 a2 b2 + b2 c 2 + c 2 a2
X bc(2a2 − bc)(b − c)2
L
≥ 0.
4a2 + b2 + c 2
.M
Without loss of generality, assume that a ≥ b ≥ c. Then, it suffices to prove that
D
c(2b2 − ca)(c − a)2 b(2c 2 − a b)(a − b)2
+ ≥ 0.
4b2 + c 2 + a2 4c 2 + a2 + b2
A
Since PI
2b2 − ca ≥ c(b + c) − ca = c(b + c − a) ≥ 0
M
and
LY
(2b2 − ca) + (2c 2 − a b) = 2(b2 + c 2 ) − a(b + c) ≥ (b + c)2 − a(b + c)

= (b + c)(b + c − a) ≥ 0,
O

.M
c(a − c)2 b(a − b)2

≥ .
4b2 + c 2 + a2 4c 2 + a2 + b2
W
This follows by multiplying the inequalities

W
c 2 (a − c)2 ≥ b2 (a − b)2
W
and
b c
≥ 2 .
4b2 + c 2 + a2 4c + a2 + b2
These inequalities are true, since
c(a − c) − b(a − b) = (b − c)(b + c − a) ≥ 0,
b(4c 2 + a2 + b2 ) − c(4b2 + c 2 + a2 ) = (b − c)[(b − c)2 + a2 − bc] ≥ 0.

The equality occurs for an equilateral triangle, and for a degenerate triangle with a = 0
1 1 1 9
+ 2 + 2 ≤ .
b2 +c 2 c +a 2 a +b 2 2(a b + bc + ca)
Solution. Apply the SOS method. Write the inequality as

X 3 a b + bc + ca
− ≥ 0,
2 b2 + c 2
L
X 3(b2 + c 2 ) − 2(a b + bc + ca)
≥ 0,
.M
b2 + c 2
X 3b(b − a) + 3c(c − a) + c(a − b) + b(a − c)
≥ 0,
D
b2 + c 2
A
X (a − b)(c − 3b) + (a − c)(b − 3c)
≥ 0,
b2 + c 2 PI
X (a − b)(c − 3b) X (b − a)(c − 3a)
+ ≥ 0,
M
b2 + c 2 c 2 + a2
X
LY
(a2 + b2 )(a − b)2 (ca + c b + 3c 2 − 3a b) ≥ 0.
Without loss of generality, assume that a ≥ b ≥ c. Since a b + ac + 3a2 − 3bc > 0, it

O
suffices to prove that

.M
(a2 + b2 )(a − b)2 (ca + c b + 3c 2 − 3a b) + (a2 + c 2 )(a − c)2 (a b + bc + 3b2 − 3ac) ≥ 0,

W
or, equivalently,
(a2 + c 2 )(a − c)2 (a b + bc + 3b2 − 3ac) ≥ (a2 + b2 )(a − b)2 (3a b − 3c 2 − ca − c b).
W
W
Since
bc + 3b2 bc + 3b2

2
a b + bc + 3b − 3ac = a + b − 3c ≥ a + b − 3c
a b+c
a(b − c)(4b + 3c)

= ≥0
b+c
and
(a b + bc + 3b2 − 3ac) − (3a b − 3c 2 − ca − c b) = 3(b2 + c 2 ) + 2bc − 2a(b + c)
≥ 3(b2 + c 2 ) + 2bc − 2(b + c)2 = (b − c)2 ≥ 0,

(a2 + c 2 )(a − c)2 ≥ (a2 + b2 )(a − b)2 ).
This is true, since is equivalent to (b − c)A ≥ 0, where
A = 2a3 − 2a2 (b + c) + 2a(b2 + bc + c 2 ) − (b + c)(b2 + c 2 )
b + c 2 a(3b2 + 2bc + 3c 2 )

= 2a a − + − (b + c)(b2 + c 2 )
2 2
b(3b2 + 2bc + 3c 2 )
− (b + c)(b2 + c 2 )
L
≥
2
.M
(b − c)(b2 + bc + 2c 2 )
= ≥ 0.
2
D
The equality occurs for an equilateral triangle, and for a degenerate triangle with a = 0
A
PI
M
LY
a + b b + c c + a

O
(a) a − b + b − c + c − a > 5;

.M
a + b2 b2 + c 2 c 2 + a2
2
(b) a2 − b2 + b2 − c 2 + c 2 − a2 ≥ 3.

W

W
Solution. Since the inequalities are symmetric, we consider that a > b > c.
W
(a) Let x = a − c and y = b − c. From a > b > c and a ≤ b + c, it follows that

x > y > 0 and c ≥ x − y. We have
a+b b+c c+a 2c + x + y 2c + y 2c + x

+ + = + −
a−b b−c c−a x−y y x
1 1 1 x+y 2c x + y

= 2c + − + > +
x−y y x x−y y x−y
2(x − y) x + y x−y

y
≥ + =2 + + 1 ≥ 5.
y x−y y x−y
(b) We will show that
a2 + b2 b2 + c 2 c 2 + a2
+ + ≥ 3;
a2 − b2 b2 − c 2 c 2 − a2
that is,
b2 c2 a2
+ ≥ .
a2 − b2 b2 − c 2 a2 − c 2
Since
a2 (b + c)2
≤ ,
a2 − c 2 a2 − c 2
L
b2 c2 (b + c)2
.M
+ ≥ .
a2 − b2 b2 − c 2 a2 − c 2
This is equivalent to each of the following inequalities:
D
1 1 1 1 2bc

A
2 2
b 2 2
− 2 2
+c 2 2
− 2 2
≥ 2 ,
a −b a −c b −c a −c a − c2
PI
b2 (b2 − c 2 ) c 2 (a2 − b2 )
+ ≥ 2bc,
M
a2 − b2 b2 − c 2
[b(b2 − c 2 ) − c(a2 − b2 )]2 ≥ 0.
LY
This completes the proof. If a > b > c, then the equality holds for a degenerate triangle
O
with a = b + c and b/c = x 1 , where x 1 ≈ 1.5321 is the positive root of the equation
x 3 − 3x − 1 = 0.
.M
W

W
b+c c+a a+b

a b c
+ + +3≥6 + + .
a b c b+c c+a a+b
W

X b+c X
2a

−6≥3 −3 .
a b+c
Since
X b+c X b c X (b − c)2
−6= + −6=
a c b bc
and X 2a X 2a − b − c X a − b X a − c
−3= = +
b+c b+c b+c b+c
X a−b X b − a X (a − b)2 X (b − c)2

= + = = ,
b+c c+a (b + c)(c + a) (c + a)(a + b)
X
a(b + c)Sa (b − c)2 ≥ 0,
where
Sa = a(a + b + c) − 2bc.
Without loss of generality, assume that a ≥ b ≥ c. Since Sa > 0,
L
S b = b(a + b + c) − 2ca = (b − c)(a + b + c) + c(b + c − a) ≥ 0
.M
and X
a(b + c)Sa (b − c)2 ≥ b(c + a)S b (c − a)2 + c(a + b)Sc (a − b)2
D
≥ (a − b)2 [b(c + a)S b + c(a + b)Sc ],
A
it suffices to prove that PI
b(c + a)S b + c(a + b)Sc ≥ 0.
M
LY
(a + b + c)[a(b2 + c 2 ) + bc(b + c)] ≥ 2a bc(2a + b + c),

O
a(a + b + c)(b − c)2 + (a + b + c)[2a bc + bc(b + c)] ≥ 2a bc(2a + b + c),

.M
a(a + b + c)(b − c)2 + bc(2a + b + c)(b + c − a) ≥ 0.

Since the last inequality is true, the proof is completed. The equality occurs for an
W
equilateral triangle, and for a degenerate triangle with a/2 = b = c (or any cyclic
permutation).
W
W
X 3a(b + c) − 2bc 3
≥ .
(b + c)(2a + b + c) 2

X 3a(b + c) − 2bc 1

− ≥ 0,
(b + c)(2a + b + c) 2
X 4a(b + c) − 6bc − b2 − c 2
≥ 0,
(b + c)(2a + b + c)
X b(a − b) + c(a − c) + 3b(a − c) + 3c(a − b)
≥ 0,
(b + c)(2a + b + c)
X (a − b)(b + 3c) + (a − c)(c + 3b)
≥ 0,
(b + c)(2a + b + c)
X (a − b)(b + 3c) X (b − a)(a + 3c)
+ ≥ 0,
(b + c)(2a + b + c) (c + a)(2b + c + a)
b + 3c a + 3c
X
(a − b) − ≥ 0,
(b + c)(2a + b + c) (c + a)(2b + c + a)
L
.M
X
(a − b)(b − c)(c − a) (a2 − b2 )(a + b + 2c) ≥ 0.
Since
D
X
(a2 − b2 )(a + b + 2c) = (a − b)(b − c)(c − a),
A
the conclusion follows. The equality holds for a = b, or b = c, or c = a.
PI
M
LY
X a(b + c) − 2bc
≥ 0.
O
(b + c)(3a + b + c)
.M
Solution. We apply the SOS method. Since

W
X a(b + c) − 2bc X b(a − c) + c(a − b)

=
(b + c)(3a + b + c) (b + c)(3a + b + c)
W
X c(b − a) X c(a − b)
W
= +
(c + a)(3b + c + a) (b + c)(3a + b + c)
X c(a + b − c)(a − b)2
= ,
(b + c)(c + a)(3a + b + c)(3b + c + a)
X
c(a + b)(3c + a + b)(a + b − c)(a − b)2 ≥ 0.
Without loss of generality, assume that a ≥ b ≥ c. Since a + b − c ≥ 0, it suffices to show

that
b(c + a)(3b + c + a)(c + a − b)(a − c)2 ≥ a(b + c)(3a + b + c)(a − b − c)(b − c)2 .
This is true since

c + a − b ≥ a − b − c,
b2 (a − c)2 ≥ a2 (b − c)2 ,
c + a ≥ b + c,
a(3b + c + a) ≥ b(3a + b + c).
L
P 1.119. Let a, b, c be positive real numbers such that a2 + b2 + c 2 ≥ 3. Prove that
.M
a5 − a2 b5 − b2 c5 − c2
+ + ≥ 0.
a5 + b2 + c 2 b5 + c 2 + a2 c 5 + a2 + b2
D
A
Solution. The inequality is equivalent to
1 1 1
PI
3
M
+ 5 + 5 ≤ 2 .
a5 +b +c
2 2 b +c +a
2 2 c +a +b
2 2 a + b2 + c 2
LY
Setting a = t x, b = t y and c = tz, where t > 0 and x, y, z > 0 such that x 2 + y 2 +z 2 = 3,

the condition a2 + b2 + c 2 ≥ 3 implies t ≥ 1, and the inequality becomes
O
1 1 1
+ + ≤ 1.
.M
t 3 x 5 + y 2 + z2 t 3 y 5 + z2 + x 2 t 3z5 + x 2 + y 2
We see that it suffices to prove this inequality for t = 1, when it becomes
W
1 1 1
+ + ≤ 1.
x 5 − x 2 + 3 y 5 − y 2 + 3 z5 − z2 + 3
W
Without loss of generality, assume that x ≥ y ≥ z. There are two cases to consider.
W
p
Case 1: z ≤ y ≤ x ≤ 2. The desired inequality follows by adding the inequalities
1 3 − x2 1 3 − y2 1 3 − z2
≤ , ≤ , ≤ .
x5 − x2 + 3 6 y5 − y2 + 3 6 z5 − z2 + 3 6
We have
1 3 − x2 (x − 1)2 (x 5 + 2x 4 − 3x 2 − 6x − 3)
− = ≤ 0,
x5 − x2 + 3 6 6(x 5 − x 2 + 3)
since
6 3
x 5 + 2x 4 − 3x 2 − 6x − 3 = x 2 (x 3 + 2x 2 − 3 − − 2)
x x
p p 3 p 1
≤ x 2 (2 2 + 4 − 3 − 3 2 − ) = −x 2 ( 2 + ) < 0.
2 2
p
Case 2: x > 2. From x 2 + y 2 + z 2 = 3, it follows that y 2 + z 2 < 1. Since
1 1 1 1
< p < p <
x5 − x + 3 (2 2 − 1)x + 3 2(2 2 − 1) + 3 6
2 2
and
1 1 1 1
+ 5 < + ,
y5 − y +3 z −z +3 3− y
2 2 2 3 − z2
L
.M
1 1 5
2
+ 2
≤ .
3− y 3−z 6
D
Indeed, we have
A
1
+
1
−
3 − y 2 3 − z2 6
=
PI
5 9( y 2 + z 2 − 1) − 5 y 2 z 2
6(3 − y 2 )(3 − z 2 )
< 0,
M
which completes the proof. The equality occurs for a = b = c = 1.
LY
p
3
Remark. Since a bc ≥ 1 involves a2 + b2 + c 2 ≥ 3 a2 b2 c 2 ≥ 3, the original inequality
is also true for a bc ≥ 1, which is a problem from IMO-2005 (by Hojoo Lee). A proof of
O
this inequality is the following:

.M
X a5 − a2 X a3 − 1
≥
a5 + b2 + c 2 a(a2 + b2 + c 2 )
W
1 X
2 1 1 X
= (a − ) ≥ (a2 − bc)
W
a2 + b2 + c 2 a a2 + b2 + c 2
W
1 X
= (a − b)2 ≥ 0.
2(a2 + b2 + c 2 )
P 1.120. Let a, b, c be positive real numbers such that a2 + b2 + c 2 = a3 + b3 + c 3 . Prove

that
a2 b2 c2 3
+ + ≥ .
b+c c+a a+b 2
(Pham Huu Duc, 2008)
First Solution. By the Cauchy-Schwarz inequality, we have

( a3 )2 ( a3 )( a2 )
P P P
X a2
≥P = P P .
b+c a4 (b + c) ( a3 )( a b) − a bc a2
P
Therefore, it is enough to show that

X X X X X
2( a3 )( a2 ) + 3a bc a2 ≥ 3( a3 )( a b).
Write this inequality as follows
X X X X X
3( a3 )( a2 − a b) − ( a3 − 3a bc) a2 ≥ 0,
L
X X X X X X X
3( a3 )( a2 − a b) − ( a)( a2 − a b) a2 ≥ 0,
.M
X X X X X
( a2 − a b)[3 a3 − ( a)( a2 )] ≥ 0.
The last inequality is true, since
D
X X X
a2 − a b) = (a − b)2 ≥ 0
A
2(
and
X X X X X
PI X
a3 − ( a2 ) = (a3 + b3 ) − a b(a + b) = (a + b)(a − b)2 ≥ 0.
M
3 a)(
The equality occurs for a = b = c = 1.
LY
Second Solution. Write the inequality in the homogeneous form A ≥ B, where

O
X a2 X 3(a3 + b3 + c 3 ) X
A=2 − a, B= − a.
b+c a2 + b2 + c 2
.M
Since
X a(a − b) + a(a − c) X a(a − b) X b(b − a)
A= = +
W
b+c b+c c+a

X (a − b)2
W
= (a + b + c)
(b + c)(c + a)
W
and
(a + b3 ) − a b(a + b) (a + b)(a − b)2
P 3 P P
B= = ,
a2 + b2 + c 2 a2 + b2 + c 2
X a + b + c a+b

− (a − b)2 ≥ 0,
(b + c)(c + a) a2 + b2 + c 2
X (a − b)2
(a3 + b3 + c 3 − 2a bc) ≥ 0.
(b + c)(c + a)
Since a3 + b3 + c 3 ≥ 3a bc, the conclusion follows.
P 1.121. If a, b, c ∈ [0, 1], then
a b c
(a) + + ≤ 1;
bc + 2 ca + 2 a b + 2
ab bc ca
(b) + + ≤ 1.
2bc + 1 2ca + 1 2a b + 1
Solution. (a) It suffices to show that
a b c
+ + ≤ 1,
L
a bc + 2 a bc + 2 a bc + 2
.M
a bc + 2 ≥ a + b + c.
D
We have
a bc + 2 − a − b − c = (1 − b)(1 − c) + (1 − a)(1 − bc) ≥ 0.
A
PI
permutation).
M
(b) It suffices to prove that
LY
ab bc ca
+ + ≤ 1;
2a bc + 1 2a bc + 1 2a bc + 1
O
that is,
2a bc + 1 ≥ a b + bc + ca.
.M
Since
a + b + c − (a b + bc + ca) = a(1 − b) + b(1 − c) + c(1 − a) ≥ 0,
W
we have
W
2a bc + 1 − a b − bc − ca ≥ 2a bc + 1 − a − b − c
W
= (1 − b)(1 − c) + (1 − a)(1 − bc) ≥ 0.
permutation).

1 1 1

5(1 − a b − bc − ca) + + + 9 ≥ 0.
1 − a b 1 − bc 1 − ca
5a(b + c) 5b(c + a) 5c(a + b)

24 − − − ≥ 0.
1 − bc 1 − ca 1 − ab
Since
4(1 − bc) ≥ 4 − (b + c)2 = (a + b + c)2 − (b + c)2 = a(a + 2b + 2c),

b+c c+a a+b

6−5 − − ≥ 0,
a + 2b + 2c b + 2c + 2a c + 2a + 2b
L
.M
X b+c

5 1− ≥ 9,
a + 2b + 2c
D
X 1
5(a + b + c)
A
≥ 9,
a + 2b + 2c
X
(a + 2b + 2c)
X PI
1

≥ 9.
a + 2b + 2c
M
The last inequality follows immediately from the AM-HM inequality. The equality holds
LY
for a = b = c = 2/3.
O
.M
2 − a2 2 − b2 2 − c2
+ +
W
≤ 3.
2 − bc 2 − ca 2 − a b
W

W
First Solution. Write the inequality as follows

X 2 − a2

1− ≥ 0,
2 − bc
X a2 − bc
≥ 0,
2 − bc
X
(a2 − bc)(2 − ca)(2 − a b) ≥ 0,
X
(a2 − bc)[4 − 2a(b + c) + a2 bc] ≥ 0,
X X X
4 (a2 − bc) − 2 a(b + c)(a2 − bc) + a bc a(a2 − bc) ≥ 0.
By virtue of the AM-GM inequality,

X
a(a2 − bc) = a3 + b3 + c 3 − 3a bc ≥ 0.

X X
2 (a2 − bc) ≥ a(b + c)(a2 − bc).
Indeed, we have
X X X
a(b + c)(a2 − bc) = a3 (b + c) − a bc (b + c)
X X X
= a(b3 + c 3 ) − a bc (b + c) = a(b + c)(b − c)2
L
.M
X a + (b + c) 2 X X
≤ (b − c)2 = (b − c)2 = 2 (a2 − bc).
2
D
The equality holds for a = b = c = 2/3, and for a = 0 and b = c = 1 (or any cyclic
A
permutation).
PI
Second Solution. We apply the SOS method. Write the inequality as follows
X a2 − bc
M
≥ 0,
2 − bc
LY
X (a − b)(a + c) + (a − c)(a + b)
≥ 0,
2 − bc
O
X (a − b)(a + c) X (b − a)(b + c)
+ ≥ 0,
2 − bc 2 − ca
.M
X (a − b)2 [2 − c(a + b) − c 2 ]
≥ 0,
(2 − bc)(2 − ca)
W
X
(a − b)2 (2 − a b)(1 − c) ≥ 0.
W
Assuming that a ≥ b ≥ c, it suffices to prove that

W
(b − c)2 (2 − bc)(1 − a) + (c − a)2 (2 − ca)(1 − b) ≥ 0.
Since 2(1 − b) = a − b + c ≥ 0 and (c − a)2 ≥ (b − c)2 , it suffices to show that
(2 − bc)(1 − a) + (2 − ca)(1 − b) ≥ 0.
We have
(2 − bc)(1 − a) + (2 − ca)(1 − b) = 4 − 2(a + b) − c(a + b) + 2a bc
(a + b) + (2 + c) 2

≥ 4 − (a + b)(2 + c) ≥ 4 − = 0.
2
3 + 5a2 3 + 5b2 3 + 5c 2
+ + ≥ 12.
3 − bc 3 − ca 3 − ab

X 3 + 5a2
− 4 ≥ 0,
3 − bc
X 5a2 + 4bc − 9
L
≥ 0,
3 − bc
.M
X 4a2 − b2 − c 2 − 2a b + 2bc − 2ca
≥ 0,
3 − bc
D
X (2a2 − b2 − c 2 ) + 2(a − b)(a − c)
A
≥ 0,
3 − bc
PI
X [(a − b)(a + b) + (a − b)(a − c)] + [(a − c)(a + c) + (a − c)(a − b)]
≥ 0,
3 − bc
M
X (a − b)(2a + b − c) + (a − c)(2a + c − b)
≥ 0,
LY
3 − bc
X (a − b)(2a + b − c) X (b − a)(2b + a − c)
+
O
≥ 0,
3 − bc 3 − ca
X (a − b)2 [3 − 2c(a + b) + c 2 ]
.M
≥ 0,
(3 − bc)(3 − ca)
X (a − b)2 (c − 1)2
W
≥ 0.
(3 − bc)(3 − ca)
W
permutation).
W
P 1.125. Let a, b, c be nonnegative real numbers such that a + b + c = 2. If

−1 7
≤m≤ ,
7 8
then
a2 + m b2 + m c2 + m 3(4 + 9m)
+ + ≥ .
3 − 2bc 3 − 2ca 3 − 2a b 19

X a2 + m 4 + 9m
− ≥ 0,
3 − 2bc 19
X 19a2 + 2(4 + 9m)bc − 12 − 8m
≥ 0.
3 − 2bc
Since
19a2 + 2(4 + 9m)bc − 12 − 8m = 19a2 + 2(4 + 9m)bc − (3 + 2m)(a + b + c)2
L
= (16 − 2m)a2 − (3 + 2m)(b2 + c 2 + 2a b + 2ac) + 2(1 + 7m)bc
.M
= (3 + 2m)(2a2 − b2 − c 2 ) + 2(5 − 3m)(a2 + bc − a b − ac) + (4 − 10m)(a b + ac − 2bc)
= (3 + 2m)(a2 − b2 ) + (5 − 3m)(a − b)(a − c) + (4 − 10m)c(a − b)
D
+(3 + 2m)(a2 − c 2 ) + (5 − 3m)(a − c)(a − b) + (4 − 10m)b(a − c)
A
= (a − b)B + (a − c)C,
PI
where
M
B = (8 − m)a + (3 + 2m)b − (1 + 7m)c,
LY
C = (8 − m)a + (3 + 2m)c − (1 + 7m)b,

O
B1 + C1 ≥ 0,
.M
where X (a − b)[(8 − m)a + (3 + 2m)b − (1 + 7m)c]

B1 = ,
3 − 2bc
W
X (b − a)[(8 − m)b + (3 + 2m)a − (1 + 7m)c]

C1 = .
W
3 − 2ca
We have
W
X (a − b)2 Ec
B1 + C1 = ,
(3 − 2bc)(3 − 2ca)
where
Ec =3(5 − 3m) − 2(8 − m)c(a + b) + 2(1 + 7m)c 2

=6(2m + 3)c 2 − 4(8 − m)c + 3(5 − 3m)
2
8−m (1 + 7m)(7 − 8m)

=6(2m + 3) c − + .
3(2m + 3) 3(2m + 3)
Since Ec ≥ 0 for −1/7 ≤ m ≤ 7/8, we get B1 + C1 ≥ 0. Thus, the proof is completed.

The equality holds for a = b = c = 2/3. When m = −1/7, the equality holds for
a = b = c = 2/3, and for a = 0 and b = c = 1 (or any cyclic permutation). When

m = 7/8, the equality holds for a = b = c = 2/3, and for a = 1 and b = c = 1/2 (or any
Remark. The inequalities in P 1.124 and P 1.125 are particular cases (k = 3 and k =
8/3, respectively) of the following more general result:
• Let a, b, c be nonnegative real numbers such that a + b + c = 3. For 0 < k ≤ 3 and
m1 ≤ m ≤ m2 , where
3

 −∞,

 0<k≤
2
L
m1 = (3 − k)(4 − k) 3 ,
.M

 , < k ≤ 3
 2(3 − 2k) 2
D
p
36 − 4k − k2 + 4(9 − k) 3(3 − k)
m2 = ,
A
72 + k
then
a2 + mbc b2 + mca c 2 + ma b
+ + ≥
PI
3(1 + m)
,
M
9 − k bc 9 − kca 9 − ka b 9−k
LY
with equality for a = b = c = 1. When m = m1 and 3/2 < k ≤ 3, the equality holds also
for
O
3
a = 0, b = c = .
2
.M
When m = m2 , the equality holds also for

p p
3k − 6 + 2 3(3 − k) 3 − 3(3 − k)
W
a= , b=c= .
k k
W
W
47 − 7a2 47 − 7b2 47 − 7c 2
+ + ≥ 60.
1 + bc 1 + ca 1 + ab

X 47 − 7a2
− 20 ≥ 0,
1 + bc
X 27 − 7a2 − 20bc
≥ 0,
1 + bc
X 3(a + b + c)2 − 7a2 − 20bc
≥ 0,
1 + bc
X −3(2a2 − b2 − c 2 ) + 2(a − b)(a − c) + 8(a b − 2bc + ca)
≥ 0,
1 + bc
X −3(a − b)(a + b) + (a − b)(a − c) + 8c(a − b)
+
1 + bc
X −3(a − c)(a + c) + (a − c)(a − b) + 8b(a − c)
+ ≥ 0,
1 + bc
L
X (a − b)(−2a − 3b + 7c) X (a − c)(−2a − 3c + 7b)
.M
+ ≥ 0,
1 + bc 1 + bc
X (a − b)(−2a − 3b + 7c) X (b − a)(−2b − 3a + 7c)
D
+ ≥ 0,
1 + bc 1 + ca
A
X (a − b)2 [1 − 2c(a + b) + 7c 2 ]
(1 + bc)(1 + ca) PI ≥ 0,
X (a − b)2 (3c − 1)2

M
≥ 0,
(1 + bc)(1 + ca)
LY
The equality holds for a = b = c = 1, and for a = 7/3 and b = c = 1/3 (or any cyclic
permutation).
O
Remark. The inequality in P 1.126 is a particular cases (k = 9) of the following more

.M
general result:
• Let a, b, c be nonnegative real numbers such that a + b + c = 3. For k > 0 and m ≥ m1 ,
W
where p
36 + 4k − k2 + 4(9 + k) 3(3 + k)

, k 6= 72
W


72 − k

m1 = ,
238
W
, k = 72


5

then
a2 + mbc b2 + mca c 2 + ma b 3(1 + m)
+ + ≤ ,
9 + k bc 9 + kca 9 + ka b 9+k
with equality for a = b = c = 1. When m = m1 , the equality holds also for

p p
3k + 6 − 2 3(3 + k) 3(3 + k) − 3
a= , b=c= .
k k
26 − 7a2 26 − 7b2 26 − 7c 2 57
+ + ≤ .
1 + bc 1 + ca 1 + ab 2

X 19 26 − 7a2
− ≥ 0,
2 1 + bc
X 14a2 + 19bc − 33
L
≥ 0,
1 + bc
.M
X 42a2 + 57bc − 11(a + b + c)2
≥ 0,
1 + bc
D
X 11(2a2 − b2 − c 2 ) + 9(a − b)(a − c) − 13(a b − 2bc + ca)
A
≥ 0,
1 + bc
PI
X 22(a − b)(a + b) + 9(a − b)(a − c) − 26c(a − b)
+
1 + bc
M
X 22(a − c)(a + c) + 9(a − c)(a − b) − 26b(a − c)
+
LY
≥ 0,
1 + bc
X (a − b)(31a + 22b − 35c) X (a − c)(31a + 22c − 35b)
O
+ ≥ 0,
1 + bc 1 + bc
.M
X (a − b)(31a + 22b − 35c) X (b − a)(31b + 22a − 35c)

+ ≥ 0,
1 + bc 1 + ca
X (a − b)2 [9 + 31c(a + b) − 35c 2 ]
W
≥ 0,
(1 + bc)(1 + ca)
W
X
(a − b)2 (1 + a b)(1 + 11c)(3 − 2c) ≥ 0.
W
Assume that a ≥ b ≥ c. Since 3 − 2c > 0, it suffices to show that
(b − c)2 (1 + bc)(1 + 11a)(3 − 2a) + (c − a)2 (1 + a b)(1 + 11b)(3 − 2b) ≥ 0;
that is,
(a − c)2 (1 + a b)(1 + 11b)(3 − 2b) ≥ (b − c)2 (1 + bc)(1 + 11a)(2a − 3).
Since 3 − 2b = a − b + c ≥ 0, we get this inequality by multiplying the inequalities
3 − 2b ≥ 2a − 3,
a(1 + a b) ≥ b(1 + bc),

a(1 + 11b) ≥ b(1 + 11a),
b2 (a − c)2 ≥ a2 (b − c)2 .
permutation).
Remark. The inequalities in P 1.127 is a particular cases (k = 9) of the following more

general result:
• Let a, b, c be nonnegative real numbers such that a + b + c = 3. For k > 0 and m ≤ m2 ,
where
(3 + k)(4 + k)
L
m2 = ,
2(3 + 2k)
.M
then
a2 + mbc b2 + mca c 2 + ma b 3(1 + m)
D
+ + ≥ ,
9 + k bc 9 + kca 9 + ka b 9+k
A
PI
with equality for a = b = c = 1. When m = m2 , the equality holds also for a = 0 and
b = c = 3/2 (or any cyclic permutation).
M
LY
O
P 1.128. Let a, b, c be nonnegative real numbers, no all are zero. Prove that
X 5a(b + c) − 6bc
.M
≤ 3.
a2 + b2 + c 2 + bc
W
First Solution. Apply the SOS method. If two of a, b, c are zero, then the inequality is
W
trivial. Consider further that a2 + b2 + c 2 = 1, a ≥ b ≥ c, b > 0, and write the inequality

W
as follows
X 5a(b + c) − 6bc

1− 2 ≥ 0,
a + b2 + c 2 + bc
X a2 + b2 + c 2 − 5a(b + c) + 7bc
≥ 0,
a2 + b2 + c 2 + bc
X (7b + 2c − a)(c − a) − (7c + 2b − a)(a − b)
≥ 0,
1 + bc
X (7c + 2a − b)(a − b) X (7c + 2b − a)(a − b)
− ≥ 0,
1 + ca 1 + bc
X
(a − b)2 (1 + a b)(3 + ac + bc − 7c 2 ] ≥ 0.
Since
3 + ac + bc − 7c 2 = 3a2 + 3b2 + ac + bc − 4c 2 > 0,
(1 + bc)(3 + a b + ac − 7a2 )(b − c)2 + (1 + ac)(3 + a b + bc − 7b2 )(a − c)2 ≥ 0.
Since
3 + a b + ac − 7b2 = 3(a2 − b2 ) + 3c 2 + b(a − b) + bc ≥ 0
and 1 + ac ≥ 1 + bc, it is enough to show that
(3 + a b + ac − 7a2 )(b − c)2 + (3 + a b + bc − 7b2 )(a − c)2 ≥ 0.
L
From b(a − c) ≥ a(b − c) ≥ 0, we get b2 (a − c)2 ≥ a2 (b − c)2 , and hence b(a − c)2 ≥
.M
a(b − c)2 . Thus, it suffices to show that
b(3 + a b + ac − 7a2 ) + a(3 + a b + bc − 7b2 ) ≥ 0.
D
A
This is true if
b(3 + a b − 7a2 ) + a(3 + a b − 7b2 ) ≥ 0.
PI
Indeed,
M
b(3 + a b − 7a2 ) + a(3 + a b − 7b2 ) = 3(a + b)(1 − 2a b) ≥ 0,
LY
since
1 − 2a b = (a − b)2 + c 2 ≥ 0.
O
.M
Second Solution. Without loss of generality, assume that a2 + b2 +c 2 = 1 and a ≤ b ≤ c.

Setting p = a + b + c, q = a b + bc + ca and r = a bc, the inequality becomes
X 5q − 11bc
W
≤ 3,
1 + bc
W
Y X
3 (1 + bc) + (11bc − 5q)(1 + ca)(1 + a b) ≥ 0,
W
3(1 + q + pr + r 2 ) + 11(q + 2pr + 3r 2 ) − 5q(3 + 2q + pr) ≥ 0,

36r 2 + 5(5 − q)pr + 3 − q − 10q2 ≥ 0.
Since p2 − 2q = 1, the inequality has the homogeneous form
36r 2 + 5(5p2 − 11q)pr + 3(p2 − 2q)3 − q(p2 − 2q)2 − 10q2 (p2 − 2q) ≥ 0.
According to P 3.57-(a) in Volume 1, for fixed p and q, the product r = a bc is minimal

when b = c or a = 0. Therefore, since 5p2 − 11q > 0, it suffices to prove the inequality
for a = 0, and for b = c = 1. For a = 0, the original inequality becomes
−6bc 10bc
+ 2 ≤ 3,
b2 + c + bc b + c 2
2
which reduces to
(b − c)2 (3b2 + 5bc + 3b2 ) ≥ 0,
while for b = c = 1, we get
10a − 6 5−a
+2 2 ≤ 3,
a2 + 3 a +a+2
a(3a + 1)(a − 1)2 ≥ 0.
Remark. Similarly, we can prove the following generalization:
• Let a, b, c be nonnegative real numbers, no all are zero. If k > 0, then
L
X (2k + 3)a(b + c) + (k + 2)(k − 3)bc
.M
≤ 3k,
a2 + b2 + c 2 + k bc
D
with equality for a = b = c, and for a = 0 and b = c (or any cyclic permutation).
A
PI
P 1.129. Let a, b, c be nonnegative real numbers, no two of which are zero, and let
M
a2 + b2 + c 2
x=
LY
.
a b + bc + ca
Prove that
O
a b c 1 1
+ + + ≥x+ ;
.M
(a)
b+c c+a a+b 2 x
4

a b c
(b) 6 + + ≥ 5x + ;
W
b+c c+a a+b x

W
3 1 1

a b c
(c) + + − ≥ x− .
b+c c+a a+b 2 3 x
W

Solution. We will prove the more general inequality
2a 2b 2c 2(1 − k)
+ + + 1 − 3k ≥ (2 − k)x + ,
b+c c+a a+b x
p
where 0 ≤ k ≤ (21+6 6)/25. For k = 0, k = 1/3 and k = 4/3, we get the inequalities in
(a), (b) and (c), respectively. Let p = a+b+c and q = a b+bc+ca. Since x = (p2 −2q)/q,
a b c
+ + ≥ f (p, q),
b+c c+a a+b
X a
+ 1 ≥ 3 + f (p, q),
b+c
p(p2 + q)
≥ 3 + f (p, q).
pq − a bc
According to P 3.57-(a) in Volume 1, for fixed p and q, the product a bc is minimal when
b = c or a = 0. Therefore, it suffices to prove the inequality for a = 0, and for b = c = 1.
For a = 0, using the substitution y = b/c + c/b, the desired inequality becomes
2(1 − k)
2 y + 1 − 3k ≥ (2 − k) y + ,
y
L
( y − 2)[k( y − 1) + 1]
.M
≥ 0.
y
Since y ≥ 2, this inequality is clearly true. For b = c = 1, the desired inequality becomes
D
4 (2 − k)(a2 + 2) 2(1 − k)(2a + 1)
A
a+ + 1 − 3k ≥ + ,
a+1 2a + 1 a2 + 2
PI
M
a(a − 1)2 [ka2 + 3(1 − k)a + 6 − 4k] ≥ 0.
LY
p
For 0 ≤ k ≤ 1, this is obvious, and for 1 < k ≤ (21 + 6 6)/25, we have
O
Æ
ka2 + 3(1 − k)a + 6 − 4k ≥ [2 k(6 − 4k) + 3(1 − k)]a ≥ 0.
.M
W
W
P 1.130. If a, b, c are real numbers, then

W
1 1 1 9
+ 2 + 2 ≤ .
a2 + 7(b + c ) b + 7(c + a ) c + 7(a + b ) 5(a + b + c)2
2 2 2 2 2 2
Solution. Let p = a + b + c and q = a b + bc + ca. Write the inequality as f6 (a, b, c) ≥ 0,

where
Y X
f6 (a, b, c) = 9 (a2 + 7b2 + 7c 2 ) − 5p2 (b2 + 7c 2 + 7a2 )(c 2 + 7a2 + 7b2 ).
Since Y Y
(a2 + 7b2 + 7c 2 ) = [7(p2 − 2q) − 6a2 ],
f6 (a, b, c) has the highest coefficient

A = 9(−6)3 < 0.
According to P 2.75 in Volume 1, it suffices to prove the original inequality for b = c = 1,
when the inequality reduces to
(a − 1)2 (a − 4)2 ≥ 0.
Thus, the proof is completed. The equality holds for a = b = c, and for a/4 = b = c (or
L
.M
bc ca ab 3
+ 2 + 2 ≤ .
+b +c 3b + c + a 3c + a + b
D
3a2 2 2 2 2 2 2 5
(Vasile Cîrtoaje and Pham Kim Hung, 2005)
A
Solution. Write the inequality as f6 (a, b, c) ≥ 0, wherePI
Y X
f6 (a, b, c) = 3 (3a2 + b2 + c 2 ) − 5 bc(3b2 + c 2 + a2 )(3c 2 + a2 + b2 ).
M
Let p = a + b + c and q = a b + bc + ca. From
LY
Y X
f6 (a, b, c) = 3 (2a2 + p2 − 2q) − 5 bc(2b2 + p2 − 2q)(2c 2 + p2 − 2q),
O
it follows that f6 (a, b, c) has the same highest coefficient A as

X
24a2 b2 c 2 − 20 b3 c 3 ;
.M
that is,
A = 24 − 60 < 0.
W
According to P 2.75 in Volume 1, it suffices to prove the original inequality for b = c = 1,

W
when the inequality is equivalent to

(a − 1)2 (3a − 2)2 ≥ 0.
W
Thus, the proof is completed. The equality holds for a = b = c, and for 3a/2 = b = c
Remark. The inequality in P 1.131 is a particular case (k = 3) of the following more
general result (Vasile Cîrtoaje, 2008):
• Let a, b, c be real numbers. If k > 1, then
X k(k − 3)a2 + 2(k − 1)bc 3(k + 1)(k − 2)
≤ ,
ka2+ b2 + c2 k+2
with equality for a = b = c, and for ka/2 = b = c (or any cyclic permutation).
P 1.132. If a, b, c are real numbers such that a + b + c = 3, then
1 1 1 3
(a) + + ≤ ;
2+ b +c
2 2 2+c +a
2 2 2+a + b
2 2 4
1 1 1 1
(b) + + ≤ .
8 + 5(b + c ) 8 + 5(c + a ) 8 + 5(a + b ) 6
2 2 2 2 2 2
(Vasile Cîrtoaje, 2006, 2009)
Solution. (a) Rewrite the inequality as follows

X 1 1 3 3

L
− ≤ − ,
2+ b +c
2 2 2 4 2
.M
b2 + c 2
X 3
≥ .
2+ b +c
2 2 2
D
A
Pp 2
X b2 + c 2
2 + b2 + c 2
≥ P
b2 + c 2
(2 + b2 + c 2 )
=
a + PI
P 2 Pp
(a2 + b2 )(a2 + c 2 )
3 + a2
P .
M
LY
XÆ X
2 (a2 + b2 )(a2 + c 2 ) ≥ a2 + 9.
O
Indeed, applying again the Cauchy-Schwarz inequality, we get

.M
XÆ X X X 2 X
2 (a2 + b2 )(a2 + c 2 ) ≥ 2 (a2 + bc) = a2 + a = a2 + 9.
W

(b) Denote p = a + b + c and q = a b + bc + ca, we write the inequality in the
W
homogeneous form
W
1 1 1 1
+ 2 + 2 ≤ 2,
8p2 + 45(b + c ) 8p + 45(c + a ) 8p + 45(a + b ) 6p
2 2 2 2 2 2
which is equivalent to f6 (a, b, c) ≥ 0, where

Y
f6 (a, b, c) = (53p2 − 90q − 45a2 )
X
−6p2 (53p2 − 90q − 45b2 )(53p2 − 90q − 45c 2 ).
Clearly, f6 (a, b, c) has the highest coefficient
A = (−45)3 < 0.
By P 2.75 in Volume 1, it suffices to prove the original inequality for b = c. In this case,
(a − 1)2 (a − 13)2 ≥ 0.
The equality holds for a = b = c = 1, and for a = 13/5 and b = c = 1/5 (or any cyclic
permutation).
L
.M
(a + b)(a + c) (b + c)(b + a) (c + a)(c + b) 4
+ + ≤ .
a2 + 4(b2 + c 2 ) b2 + 4(c 2 + a2 ) c 2 + 4(a2 + b2 ) 3
D
A
where Y
PI
f6 (a, b, c) = 4 (a2 + 4b2 + 4c 2 )
M
X
LY
−3 (a + b)(a + c)(b2 + 4c 2 + 4a2 )(c 2 + 4a2 + 4b2 )

Y X
=4 (4p2 − 8q − 3a2 ) − 3 (a2 + q)(4p2 − 8q − 3b2 )(4p2 − 8q − 3c 2 ).
O
Thus, f6 (a, b, c) has the highest coefficient

.M
A = 4(−3)3 − 34 < 0.
W
By P 2.75 in Volume 1, it suffices to prove the original inequality for b = c = 1, when

W
(a − 1)2 (2a − 7)2 ≥ 0.

W
The equality holds for a = b = c, and for 2a/7 = b = c (or any cyclic permutation).
X 1 1
≤ .
(b + c)(7a + b + c) 2(a b + bc + ca)


X 4(a b + bc + ca)

1− ≥ 1,
(b + c)(7a + b + c)
X (b − c)2 + 3a(b + c)
≥ 1.
(b + c)(7a + b + c)
X (b − c)2 + 3a(b + c) 4(a + b + c)4
≥P .
(b + c)(7a + b + c) [(b − c)2 + 3a(b + c)](b + c)(7a + b + c)
L
.M
X
4(a + b + c)4 ≥ (b2 + c 2 − 2bc + 3ca + 3a b)(b + c)(7a + b + c).
D
X X X X
a4 + a bc a+3 a b(a2 + b2 ) − 8 a2 b2 ≥ 0,
A
X
a4 + a bc
X
a−
X
PI
a b(a2 + b2 ) + 4
X
a b(a − b)2 ≥ 0.
a + a bc a − a b(a2 + b2 ) ≥ 0 (Schur’s inequality of degree four), the
P 4 P P
M
Since
conclusion follows. The equality holds for a = b = c, and also for a = 0 and b = c (or
LY

Second Solution. Let p = a + b + c and q = a b + bc + ca. We need to prove that
O
f6 (a, b, c) ≥ 0, where Y
f6 (a, b, c) = (b + c)(7a + b + c)
.M
X
−2(a b + bc + ca) (a + b)(a + c)(7b + c + a)(7c + a + b)
W
Y X
= (p − a)(p + 6a) − 2q (p − b)(p − c)(p + 6b)(p + 6c).
W
Clearly, f6 (a, b, c) has the highest coefficient A = (−6)3 < 0. Thus, by P 3.76-(a) in
Volume 1, it suffices to prove the original inequality for b = c = 1, and for a = 0. For
W
b = c = 1, the inequality reduces to a(a − 1)2 ≥ 0, which is obviously true. For a = 0,

1 1 1 1
+ + ≤ ,
(b + c)2 c(7b + c) b(7c + b) 2bc
1 b2 + c 2 + 14bc 1
+ ≤ ,
(b + c)2 bc[7(b + c ) + 50bc] 2bc
2 2
1 x + 14 1
+ ≤ ,
x + 2 7x + 50 2
where x = b/c +c/b, x ≥ 2. This reduces to the obvious inequality (x −2)(5x +28) ≥ 0.
X 1 9
≤ .
b2 + c2 + 4a(b + c) 10(a b + bc + ca)
Solution. Let p = a + b + c and q = a b + bc + ca. We need to prove that f6 (a, b, c) ≥ 0,

where Y
f6 (a, b, c) = 9 [b2 + c 2 + 4a(b + c)]
X
−10(a b + bc + ca) [a2 + b2 + 4c(a + b)][a2 + c 2 + 4b(a + c)]
L
Y X
.M
=9 (p2 + 2q − a2 − 4bc) − 10q (p2 + 2q − c 2 − 4a b)(p2 + 2q − b2 − 4ca).
Clearly, f6 (a, b, c) has the same highest coefficient A as f (a, b, c), where
D
Y X X
f (a, b, c) = −9 (a2 + 4bc) = −9(65a2 b2 c 2 + 16a bc a3 + 4 a3 b3 );
A
that is, PI
A = −9(65 + 48 + 12) < 0.
M
Thus, by P 3.76-(a) in Volume 1, it suffices to prove the original inequality for b = c = 1,
LY
and for a = 0. For b = c = 1, the inequality reduces to a(a − 1)2 ≥ 0, which is obviously
true. For a = 0, the inequality becomes
O
1 1 1 9
+ 2 + 2 ≤ ,
+c b + 4bc c + 4bc
.M
b2 2 10bc
1 b2 + c 2 + 8bc 9
+ ≤ .
b +c 4bc(b + c ) + 17b c
W
2 2 2 2 2 2 10bc
1 x +8 9
+
W
≤ ,
x 4x + 17 10
where x = b/c +c/b, x ≥ 2. The inequality is true, since it is equivalent to (x −2)(26x +
W
85) ≥ 0. The equality holds for a = b = c, and also for a = 0 and b = c (or any cyclic
permutation).
P 1.136. If a, b, c are nonnegative real numbers such that a + b + c = 3, then
1 1 1 9
+ + ≤ .
3 − a b 3 − bc 3 − ca 2(a b + bc + ca)

First Solution. We apply the SOS method. Write the inequality as

X 3 a b + bc + ca
− ≥ 0.
2 3 − bc
X 9 − 2a(b + c) − 5bc
≥ 0,
3 − bc
X a2 + b2 + c 2 − 3bc
≥ 0.
3 − bc
Since
2(a2 + b2 + c 2 − 3bc) = 2(a2 − bc) + 2(b2 + c 2 − a b − ac) + 2(a b + ac − 2bc)
L
.M
= (a − b)(a + c) + (a − c)(a + b) − 2b(a − b) − 2c(a − c) + 2c(a − b) + 2b(a − c)
= (a − b)(a − 2b + 3c) + (a − c)(a − 2c + 3b),
D
the required inequality is equivalent to
A
X (a − b)(a − 2b + 3c) + (a − c)(a − 2c + 3b)
3 − bc
PI ≥ 0,
M
X (a − b)(a − 2b + 3c) X (b − a)(b − 2a + 3c)
+ ≥ 0,
3 − bc 3 − ca
LY
X (a − b)2 [9 − c(a + b + 3c)]

≥ 0,
(3 − bc)(3 − ca)
O
X
(a − b)2 (3 − a b)(3 + c)(3 − 2c) ≥ 0.
.M
Without loss of generality, assume that a ≥ b ≥ c. Then it suffices to prove that
(b − c)2 (3 − bc)(3 + a)(3 − 2a) + (c − a)2 (3 − ca)(3 + b)(3 − 2b) ≥ 0,

W
W
(a − c)2 (3 − ac)(3 + b)(3 − 2b) ≥ (b − c)2 (3 − bc)(a + 3)(2a − 3).

W
Since 3−2b = a− b+c ≥ 0, we can obtain this inequality by multiplying the inequalities
b2 (a − c)2 ≥ a2 (b − c)2 ,
a(3 − ac) ≥ b(3 − bc) ≥ 0,

a(3 + b)(3 − 2b) ≥ b(a + 3)(2a − 3) ≥ 0.
We have
a(3 − ac) − b(3 − bc) = (a − b)[3 − c(a + b)] = (a − b)(3 − 3c + c 2 )

≥ 3(a − b)(1 − c) ≥ 0.
Also, since a + b ≤ a + b + c = 3, we have
a(3 + b)(3 − 2b) − b(a + 3)(2a − 3) = 9(a + b) − 6a b − 2a b(a + b)
≥ 9(a + b) − 12a b ≥ 3(a + b)2 − 12a b = 3(a − b)2 ≥ 0.

permutation).
Second Solution. Let p = a + b + c and q = a b + bc + ca. We need to prove that
f6 (a, b, c) ≥ 0, where
Y X
f6 (a, b, c) = 3 (p2 − 3bc) − 2q (p2 − 2ca)(p2 − 2a b).
L
.M
Clearly, f6 (a, b, c) has the highest coefficient
A = 3(−3)3 < 0.
D
A
Thus, by P 3.76-(a) in Volume 1, it suffices to prove the original inequality for b = c,
and for a = 0. For b = c = 3 − a, the inequality reduces to
PI
a(9 − a)(a − 1)2 ≥ 0,
M
which is obviously true. For a = 0, which yields b + c = 3, the inequality can be written
LY
as
(9 − 4bc)(9 − bc) ≥ 0.
O
Indeed,
.M
(9 − 4bc)(9 − bc) = (b − c)2 (b2 + c 2 + bc) ≥ 0.

W
W

W
bc ca ab 3
+ 2 + 2 ≤ .
a2 +a+6 b + b+6 c +c+6 8

X bc 1
≤ ,
3a2 + ap + 2p 2 8
where p = a + b + c. We need to prove that f6 (a, b, c) ≥ 0, where

Y X
f6 (a, b, c) = (3a2 + ap + 2p2 ) − 8 bc(3b2 + bp + 2p2 )(3c 2 + c p + 2p2 ).
Clearly, f6 (a, b, c) has the same highest coefficient as

X
27a2 b2 c 2 − 72 b3 c 3 ,
that is,
A = 27 − 216 < 0.
Thus, by P 3.76-(a) in Volume 1, it suffices to prove that f6 (a, 1, 1) ≥ 0 and f6 (0, b, c) ≥ 0
for all a, b, c ≥ 0. Indeed, we have
f6 (a, 1, 1) = 2a(a2 + 9a + 3)(a − 1)2 (6a + 1) ≥ 0
L
and
.M
f6 (0, b, c) = 2(b − c)2 (5b2 + 5bc + 2c 2 )(2b2 + 5bc + 5c 2 ) ≥ 0.
D
permutation).
A
PI
M
1 1 1 1
LY
+ 2 + 2 ≥ .
8a2 − 2bc + 21 8b − 2ca + 21 8c − 2a b + 21 9
O
(Michael Rozenberg, 2013)

.M
Solution. Let
q = a b + bc + ca.
Write the inequality in the homogeneous form f6 (a, b, c) ≥ 0, where
W
X Y
(8b2 − 2ca + 7q)(8c 2 − 2a b + 7q) − (8a2 − 2bc + 7q).
W
f6 (a, b, c) = 3q
W
Clearly, f6 (a, b, c) has the same highest coefficient as f (a, b, c), where
Y X X
f (a, b, c) = −8 (4a2 − bc) = −8(63a2 b2 c 2 − 16 a3 b3 + 4a bc a3 );
that is,
A = −8(63 − 48 + 12) < 0.
By P 3.76-(a) in Volume 1, it suffices to prove that f6 (a, 1, 1) ≥ 0 and f6 (0, b, c) ≥ 0 for
all a, b, c ≥ 0. Indeed, we have
f6 (a, 1, 1) = 0
and
f6 (0, b, c) = 8b2 c 2 (b − c)2 ≥ 0.
The equality holds when two of a, b, c are equal.

Remark. The following identity holds:
8 (a − b)2
Q
X 9
−1= Q .
8a2 − 2bc + 21 (a2 − 2bc + 21)
P 1.139. Let a, b, c be real numbers, no two of which are zero. Prove that
L
a2 + bc b2 + ca c 2 + a b (a + b + c)2
.M
(a) + + ≥ ;
b2 + c 2 c 2 + a2 a2 + b2 a2 + b2 + c 2
a2 + 3bc b2 + 3ca c 2 + 3a b 6(a b + bc + ca)
D
(b) + 2 + 2 ≥ .
b2 + c 2 c + a2 a + b2 a2 + b2 + c 2
A
Solution. (a) Using the known inequality

PI
M
X a2 3
LY
≥
b +c
2 2 2
O
and the Cauchy-Schwarz inequality yields

X a2 + bc a2
.M
X X bc X1 bc

= + ≥ +
b2 + c 2 b2 + c 2 b2 + c 2 2 b2 + c 2
P 2
X (b + c)2 (b + c) (a + b + c)2
W
= ≥ = .
2(b2 + c 2 ) 2(b2 + c 2 ) a2 + b2 + c 2
P
W

W
(b) We have
X a2 + 3bc X a2 X 3bc 3 X 3bc
= + ≥ +
b2 + c 2 b2 + c 2 b2 + c 2 2 b2 + c 2
X1 (b + c)2

bc X
= −3 + 3 + 2 = −3 + 3
2 b + c2 2(b2 + c 2 )
P 2 P 2
3 (b + c) 3 a 6(a b + bc + ca)
≥ −3 + P = −3 + P = .
2(b + c )
2 2 a 2 a2 + b2 + c 2

P 1.140. Let a, b, c be real numbers such that a b + bc + ca ≥ 0 and no two of which are
zero. Prove that
a(b + c) b(c + a) c(a + b) 3
+ 2 + 2 ≥ .
b2 + c 2 c + a2 a + b2 10
Solution. Since the problem remains unchanged by replacing a, b, c by −a, −b, −c, it
suffices to consider the cases a, b, c ≥ 0 and a < 0, b ≥ 0, c ≥ 0.
Case 1: a, b, c ≥ 0. We have
X a(b + c) X a(b + c)
L
≥
b2 + c 2 (b + c)2
.M
X a 3 3
= ≥ > .
b+c 2 10
D
Case 2: a < 0, b ≥ 0, c ≥ 0. Replacing a by −a, we need to show that
A
b(c − a) c(b − a) a(b + c)
+ 2 − 2 ≥
3PI
a +c
2 2 a +b 2 b +c 2 10
M
for any nonnegative numbers a, b, c such that
LY
bc
a≤ .
O
b+c
.M
We show first that

b(c − a) b(c − x)
≥ 2 ,
a2 + c 2 x + c2
W
bc
where x = , x ≥ a. This is equivalent to
b+c
W
b(x − a)[(c − a)x + ac + c 2 ] ≥ 0,

W
which is true because
c 2 (a + 2b + c)
(c − a)x + ac + c 2 = ≥ 0.
b+c
Similarly, we can show that
c(b − a) c(b − x)
≥ 2 .
a +b
2 2 x + b2
In addition,
a(b + c) x(b + c)
≤ 2 .
b +c
2 2 b + c2
b(c − x) c(b − x) x(b + c) 3

+ 2 − 2 ≥ .
x +c
2 2 x +b 2 b +c 2 10
Denote
b c
p= , q= , p + q = 1.
b+c b+c
Since
b(c − x) p c(b − x) q
= , =
x +c
2 2 1 + p2 x +b
2 2 1 + q2
and
L
x(b + c) bc pq
= 2 =
.M
,
b +c
2 2 b +c 2 1 − 2pq
we need to show that
D
p q pq 3
+ − ≥ .
1+p 2 1+q 2 1 − 2pq 10
A
Since
p
+
q
=
1 + pq PI ,
1 + p2 1 + q2 2 − 2pq + p2 q2
M
LY
(pq + 2)2 (1 − 4pq) ≥ 0,

O
which is true since

.M
1 − 4pq = (p + q)2 − 4pq = (p − q)2 ≥ 0.

W
The equality holds for −2a = b = c (or any cyclic permutation).

W
W
P 1.141. If a, b, c are positive real numbers such that a bc > 1, then
1 1 4
+ ≥ .
a + b + c − 3 a bc − 1 a b + bc + ca − 3
Solution (by Vo Quoc Ba Can). By the AM-GM inequality, we have

3
p
a + b + c ≥ 3 a bc > 3,
3
p
a b + bc + ca ≥ a2 b2 c 2 > 3.
Without loss of generality, assume that a = min{a, b, c}. By the Cauchy-Schwarz in-
equality, we have
1 1 a bc − 1 p 1 2

+ a(a + b + c − 3) + ≥ a+ p .
a + b + c − 3 a bc − 1 a a
a bc − 1
a(a + b + c − 3) +
(a + 1)2
a
≥ .
4a a b + bc + ca − 3
L
Since
.M
a bc − 1 (a − 1)3
a(a + b + c − 3) + = a b + bc + ca − 3 + ,
a a
D
this inequality can be written as follows
A
(a + 1)2 (a − 1)3
4a
−1≥ PI
a(a b + bc + ca − 3)
,
M
(a − 1)2 (a − 1)3
≥ ,
a(a b + bc + ca − 3)
LY
4a
(a − 1)2 (a b + bc + ca + 1 − 4a) ≥ 0.
O
This is true since

.M
Æ
3
bc ≥ (a bc)2 > 1,
and hence
W
a b + bc + ca + 1 − 4a > a2 + 1 + a2 + 1 − 4a = 2(a − 1)2 ≥ 0.

W
The equality holds for a > 1 and b = c = 1 (or any cyclic permutation).
W
Remark. Using this inequality, we can prove P 3.84 in Volume 1, which states that
1 1 1 1

(a + b + c − 3) + + − 3 + a bc + ≥2
a b c a bc
for any positive real numbers a, b, c. This inequality is clearly true for a bc = 1. In
addition, it remains unchanged by substituting a, b, c with 1/a, 1/b, 1/c,
p respectively.
3
Therefore, it suffices to consider the case a bc > 1. Since a + b + c ≥ 3 a bc > 3, we
can write the required inequality as E ≥ 0, where
(a bc − 1)2
E = a b + bc + ca − 3a bc + .
a+ b+c−3
According to the inequality in P 1.141, we have
4 1

2
E ≥ a b + bc + ca − 3a bc + (a bc − 1) −
a b + bc + ca − 3 a bc − 1
4(a bc − 1)2
= (a b + bc + ca − 3) + − 4(a bc − 1)
a b + bc + ca − 3
v
t 4(a bc − 1)2
≥2 (a b + bc + ca − 3) · − 4(a bc − 1) = 0.
a b + bc + ca − 3
L
.M
D
X (4b2 − ac)(4c 2 − a b) 27
≤ a bc.
b+c 2
A
Solution. Since
M
X (4b2 − ac)(4c 2 − a b) X bc(16bc + a2 ) X a(b3 + c 3 )
LY
= −4
b+c b+c b+c
X bc(16bc + a2 ) X
O
= −4 a(b2 + c 2 ) + 12a bc
b + c
X a2
.M
16bc
= bc + − 4(b + c) + 12a bc
b+c b+c
X a2 (b − c)2

W
= bc −4 + 12a bc
b+c b+c
W

W
X a a2 4(b − c)2

bc − + ≥ 0,
2 b+c b+c
X bc(b − c)2 X 2a − b − c
8 ≥ a bc .
b+c b+c
In addition, since
X 2a − b − c X (a − b) + (a − c) X a−b X b−a
= = +
b+c b+c b+c c+a
X (a − b)2 X (b − c)2
= = ,
(b + c)(c + a) (c + a)(a + b)
the inequality can be restated as

X bc(b − c)2 X (b − c)2
8 ≥ a bc ,
b+c (c + a)(a + b)
X bc(b − c)2 (8a2 + 8bc + 7a b + 7ac)
≥ 0.
(a + b)(b + c)(c + a)
Since the last form is obvious, the proof is completed. The equality holds for a = b = c,
and also for a = 0 and b = c (or any cyclic permutation).
L
.M
P 1.143. Let a, b, c be nonnegative real numbers, no two of which are zero, such that
D
a + b + c = 3.
A
Prove that
a
+
b
+
c
3a + bc 3b + ca 3c + a b
2PI
≥ .
3
M
Solution. Since
LY
3a + bc = a(a + b + c) + bc = (a + b)(a + c),

O
2
.M
a(b + c) + b(c + a) + c(a + b) ≥ (a + b)(b + c)(c + a),

3
6(a b + bc + ca) ≥ 2[(a + b + c)(a b + bc + ca) − a bc],
W
2a bc ≥ 0.
W
The equality holds for a = 0, or b = 0, or c = 0.

W
P 1.144. Let a, b, c be positive real numbers such that
1 1 1

(a + b + c) + + = 10.
a b c
Prove that
19 a b c 5
≤ + + ≤ .
12 b+c c+a a+b 3
First Solution. Write the hypothesis
1 1 1

(a + b + c) + + = 10
a b c
as
b+c c+a a+b
+ + =7
a b c
and
(a + b)(b + c)(c + a) = 9a bc.
b+c c+a a+b
Using the substitutions x = , y = and z = , we need to show that
a b c
L
x + y + z = 7 and x yz = 9 involve
.M
19 1 1 1 5
≤ + + ≤ ,
12 x y z 3
D
or, equivalently,
A
19 1 x(7 − x) 5
≤ + ≤ .
PI
12 x 9 3
Clearly, x, y, z ∈ (0, 7). The left inequality is equivalent to
M
(x − 4)(2x − 3)2 ≤ 0,
LY
while the right inequality is equivalent to

O
(x − 1)(x − 3)2 ≥ 0.
.M
These inequalities are true if 1 ≤ x ≤ 4. To show that 1 ≤ x ≤ 4, from ( y + z)2 ≥ 4 yz,

we get
W
36
(7 − x)2 ≥ ,
x
W
(x − 1)(x − 4)(x − 9) ≥ 0,
W
1 ≤ x ≤ 4.
Thus, the proof is completed. The left inequality is an equality for 2a = b = c (or any
cyclic permutation), and the right inequality is an equality for a/2 = b = c (or any cyclic
permutation).
Second Solution. Due to homogeneity, assume that b + c = 2; this involves bc ≤ 1.
From the hypothesis
1 1 1

(a + b + c) + + = 10,
a b c
we get
2a(a + 2)
bc = .
9a − 2
Since
(a − 2)(2a − 1)
bc − 1 = ,
9a − 2
from the condition bc ≤ 1, we get
1
≤ a ≤ 2.
2
We have
b c a(b + c) + b2 + c 2 2a + 4 − 2bc
+ = 2 = 2
c+a a+b a + (b + c)a + bc a + 2a + bc
L
2(7a + 12a − 4) 2(7a − 2)
2
.M
= = ,
9a2 (a + 2) 9a2
D
and hence
a b c a 2(7a − 2) 9a3 + 28a − 8
+ + = + =
A
.
b+c c+a a+b 2 9a2 18a2
Thus, we need to show that PI
M
19 9a3 + 28a − 8 5
≤ ≤ .
12 18a2 3
LY
These inequalities are true, since the left inequality is equivalent to

O
(2a − 1)(3a − 4)2 ≥ 0,

.M
and the right inequality is equivalent to

W
(a − 2)(3a − 2)2 ≤ 0.
W
Remark. Similarly, we can prove the following generalization.

W
• Let a, b, c be positive real numbers such that
1 1 1 8k2

(a + b + c) + + =9+ ,
a b c 1 − k2
where k ∈ (0, 1). Then,
k2 a b c 3 k2
≤ + + − ≤ .
1+k b+c c+a a+ b 2 1−k
P 1.145. Let a, b, c be nonnegative real numbers, no two of which are zero, such that
a + b + c = 3. Prove that
9 a b c
< + + ≤ 1.
10 2a + bc 2b + ca 2c + a b
Solution. (a) Since

a 1 −bc
− = ,
2a + bc 2 2(2a + bc)
we can write the right inequality as
L
.M
X bc
≥ 1.
2a + bc
D
A
( bc)2 b c + 2a bc a
P P 2 2 P
X bc
2a + bc
≥P
bc(2a + bc)
= PI
6a bc + b2 c 2
P = 1.
M
The equality holds for a = b = c = 1, and also for a = 0, or b = 0, or c = 0.
(b) First Solution. For the nontrivial case a, b, c > 0, we can write the left inequality
LY
as X 1 9
O
> .
bc 10
2+
.M
a
Using the substitutions
W
v s v
t bc ca t ab
x= , y= , z= ,
a b c
W

W
X 1 9
>
2 + x2 10
for all positive real numbers x, y, z satisfying x y + yz + z x = 3. By expanding, the
inequality becomes X X
4 x 2 + 48 ≥ 9x 2 y 2 z 2 + 8 x 2 y 2.
Since X X X X
x2 y2 = ( x y)2 − 2x yz x = 9 − 2x yz x,
we can write the desired inequality as
X X
4 x 2 + 16x yz x ≥ 9x 2 y 2 z 2 + 24,
16x yz(x + y + z) ≥ 9x 2 y 2 z 2 + 4(2x y + 2 yz + 2z x − x 2 − y 2 − z 2 ).
Using Schur’s inequality

9x yz
≥ 2x y + 2 yz + 2z x − x 2 − y 2 − z 2 ,
x + y +z

36x yz
16x yz(x + y + z) ≥ 9x 2 y 2 z 2 + .
x + y +z
L
.M
This is true if
36
16(x + y + z) ≥ 9x yz + .
x + y +z
D
Since
A
Æ
x + y +z ≥ 3(x y + yz + z x) = 3
and
x y + yz + z x Æ
PI
3
1=
M
≥ x 2 y 2z2,
3
LY
we have
36
16(x + y + z) − 9x yz − ≥ 48 − 9x yz − 12 = 9(4 − x yz) > 0.
O
x + y +z
.M
Second Solution. As it is shown at the first solution, it suffices to show that

X 1 9
>
2+ x
W
2 10
for all positive real numbers x, y, z satisfying x y + yz + z x = 3. Rewrite this inequality
W
as
X x2 6
W
< .
2+ x 2 5
Let p and q be two positive real numbers such that
p
p + q = 3.
x2 3x 2 (p x + q x)2
= =
2 + x2 2(x y + yz + z x) + 3x 2 2x(x + y + z) + (x 2 + 2 yz)
p2 x q2 x 2
≤ + 2 .
2(x + y + z) x + 2 yz
Therefore,
X x2 X p2 x X q2 x 2 p2 2
X x2
≤ + = + q .
2 + x2 2(x + y + z) x 2 + 2 yz 2 x 2 + 2 yz
p2 X x2 6
+ q2 < .
2 x + 2 yz
2 5
We claim that
X x2
L
< 2.
x 2 + 2 yz
.M
Under this assumption, we only need to show that
D
p2 6
+ 2q2 ≤ .
A
2 5
p p
Indeed, choosing p =
4 3
and q =
3 PI p
, we have p + q = 3 and
p2 6
+ 2q2 = . To
5 5 2 5
M
x2
< 2,
P
complete the proof, we need to prove the homogeneous inequality
x 2 + 2 yz
LY
X yz 1
> .
O
x 2 + 2 yz 2
.M
By the Cauchy-Schwarz inequality, we get
( yz)2 y z + 2x yz x
P P 2 2 P
X yz 1
≥P = > .
W
x + 2 yz yz(x + 2 yz) x yz x + 2 y z
P P
2 2 2 2 2
W
W
a3 b3 c3 a3 + b3 + c 3
+ + ≤ .
2a2 + bc 2b2 + ca 2c 2 + a b a2 + b2 + c 2

X a3 a3

− ≥ 0,
a2 + b2 + c 2 2a2 + bc
X a3 (a2 + bc − b2 − c 2 )
≥ 0,
2a2 + bc
X a3 [a2 (b + c) − b3 − c 3 ]
≥ 0,
(b + c)(2a2 + bc)
X a3 b(a2 − b2 ) + a3 c(a2 − c 2 )
≥ 0,
(b + c)(2a2 + bc)
X a3 b(a2 − b2 ) X a3 c(a2 − c 2 )
+ ≥ 0,
(b + c)(2a2 + bc) (b + c)(2a2 + bc)
X a3 b(a2 − b2 ) X b3 a(b2 − a2 )
+
L
≥ 0,
(b + c)(2a2 + bc) (c + a)(2b2 + ca)
.M
X a b(a + b)(a − b)2 [2a2 b2 + c(a3 + a2 b + a b2 + b3 ) + c 2 (a2 + a b + b2 )]
≥ 0.
(b + c)(c + a)(2a2 + bc)(2b2 + ca)
D
A
tion).
PI
M
LY
a3 b3 c3 a+b+c
O
+ + ≥ .
4a + bc 4b + ca 4c + a b
2 2 2 5
.M
Solution. Assume that a ≥ b ≥ c, and write the inequality as follows

W
X a3

a
− ≥ 0,
W
4a2 + bc 5
W
X a(a2 − bc)
≥ 0,
4a2 + bc
X a[(a − b)(a + c) + (a − c)(a + b)]
≥ 0,
4a2 + bc
X a(a − b)(a + c) X a(a − c)(a + b)
+ ≥ 0,
4a2 + bc 4a2 + bc
X a(a − b)(a + c) X b(b − a)(b + c)
+ ≥ 0,
4a2 + bc 4b2 + ca
X c(a − b)2 [(a − b)2 + bc + ca − a b]
≥ 0.
(4a2 + bc)(4b2 + ca)
Clearly, it suffices to show that

X c(a − b)2 (bc + ca − a b)
≥ 0,
(4a2 + bc)(4b2 + ca)
we can be written as
X
(a − b)2 (bc + ca − a b)(4c 3 + a bc) ≥ 0.
Since ca + a b − bc > 0, it is enough to prove that
(c − a)2 (a b + bc − ca)(4b3 + a bc) + (a − b)2 (bc + ca − a b)(4c 3 + a bc) ≥ 0.
L
.M
In addition, since (c − a)2 ≥ (a − b)2 , 4b3 + a bc ≥ 4c 3 + a bc and a b + bc − ca > 0, we
only need to show that
D
(a − b)2 (a b + bc − ca)(4c 3 + a bc) + (a − b)2 (bc + ca − a b)(4c 3 + a bc) ≥ 0.
A
This is equivalent to the obvious inequality PI
a bc(a − b)2 (4c 2 + bc) ≥ 0.
M
LY

O
.M
1 1 1 3
W
+ + ≥ .
(2 + a)2 (2 + b)2 (2 + c)2 6 + a b + bc + ca
W

W

X 1 4(a + b + c)2
≥ .
(2 + a)2 (2 + a)2 (b + c)2
P

X
4(a + b + c)2 (6 + a b + bc + ca) ≥ (2 + a)2 (b + c)2 .
2p2 q − 3q2 + 3pr + 12q ≥ 6(pq + 3r),

where
According to AM-GM inequality,
Æ
2p2 q − 3q2 + 3pr + 12q ≥ 2 12q(2p2 q − 3q2 + 3pr).
Therefore, it is enough to prove the homogeneous inequality
4q(2p2 q − 3q2 + 3pr) ≥ 3(pq + 3r)2 ,
which can be written as
L
5p2 q2 ≥ 12q3 + 6pqr + 27r 2 .
.M
Since pq ≥ 9r, we have
3(5p2 q2 − 12q3 − 6pqr − 27r 2 ) ≥ 15p2 q2 − 36q3 − 2p2 q2 − p2 q2
D
= 12q2 (p2 − 3q) ≥ 0.
A
The equality holds for a = b = c = 1. PI
M
LY

O
1 1 1 3
+ + ≥ .
1 + 3a 1 + 3b 1 + 3c 3 + a bc
.M
Solution. Set
W
3
p
p = a + b + c, q = a b + bc + ca, r= a bc,
W
and write the inequality as follows

W
X
(3 + r 3 ) (1 + 3b)(1 + 3c) ≥ 3(1 + 3a)(1 + 3b)(1 + 3c),
(3 + r 3 )(3 + 6p + 9q) ≥ 3(1 + 3p + 9q + 27r 3 ),

r 3 (2p + 3q) + 2 + 3p ≥ 26r 3 .
By virtue of the AM-GM inequality, we have
p ≥ 3r, q ≥ 3r 2 .
r 3 (6r + 9r 2 ) + 2 + 9r ≥ 26r 3 ,
(r − 1)2 (9r 3 + 24r 2 + 13r + 2) ≥ 0.
P 1.150. Let a, b, c be real numbers, no two of which are zero. If 1 ≤ k ≤ 3, then
2a b 2bc 2ca

k+ 2 k + k + ≥ (k − 1)(k2 − 1).
a + b2 b2 + c 2 c 2 + a2
L
.M
Solution. If a, b, c are the same sign, then the inequality is obvious since
D
2a b 2bc 2ca

A
k+ 2 k+ 2 k+ 2 ≥ k3 > (k − 1)(k2 − 1).
a + b2 b + c2 c + a2
PI
Since the inequality remains unchanged by replacing a, b, c with −a, −b, −c, it suffices
M
to consider further that a ≤ 0, b ≥ 0, c ≥ 0. Setting −a for a, we need to show that
2a b 2bc 2ca

LY
k− 2 k+ 2 k− 2 ≥ (k − 1)(k2 − 1)
a + b2 b + c2 c + a2
O
for a, b, c ≥ 0. Since
.M
(a − b)2 (a − c)

2a b 2ca

k− 2 k − = k − 1 + k − 1 +
a + b2 c 2 + a2 a2 + b2 c 2 + a2
W
(a − b)2 (a − c)2

2
≥ (k − 1) + (k − 1) + 2 ,
a2 + b2 c + a2
W

W
(a − b)2 (a − c)2

2bc

k−1+ 2 + 2 k+ ≥ k2 − 1.
a + b2 c + a2 b2 + c 2
According to the inequality (a) from P 2.19 in Volume 3, we have
(a − b)2 (a − c)2 (b − c)2

+ ≥ .
a2 + b2 c 2 + a2 (b + c)2
(b − c)2

2bc

k−1+ k+ 2 ≥ k2 − 1,
(b + c)2 b + c2
(b − c)4 + 2(3 − k)bc(b − c)2 ≥ 0.
P 1.151. If a, b, c are non-zero and distinct real numbers, then

1 1 1 1 1 1 1 1 1

+ + +3 + + ≥4 + + .
a2 b2 c 2 (a − b)2 (b − c)2 (c − a)2 a b bc ca
L
.M
D
1 X 1 1 X 1
X X
− + 3 ≥ 3 .
a2 bc (b − c)2 bc
A
v
t X 1 X 1 X
PI
In virtue of the AM-GM inequality, it suffices to prove that
1 X 1

M
2 3 − ≥ 3 ,
a2 bc (b − c)2 bc
LY
which is true if
O
X 2
1 X 1 1 1
X X
4 − ≥3 .
a 2 bc (b − c)2 bc
.M
Rewrite this inequality as

X X X X
a2 b2 − a bc a2 − a b)2 ≥ 3(a + b + c)2 (a − b)2 (b − c)2 (c − a)2 .
W
4( a)(
W
Using the notations
p = a + b + c, q = a b + bc + ca, r = a bc,
W
and the identity
(a − b)2 (b − c)2 (c − a)2 = −27r 2 − 2(2p2 − 9q)pr + p2 q2 − 4q3 ,
4(q2 − 3pr)(p2 − 3q)2 ≥ 3p2 [−27r 2 − 2(2p2 − 9q)pr + p2 q2 − 4q3 ],
(9pr + p2 q − 6q2 )2 ≥ 0.
P 1.152. Let a, b, c be positive real numbers, and let
a b b c c a
A= + + k, B= + + k, C= + + k,
b a c b a b
where −2 < k ≤ 4. Prove that
1 1 1 1 4
+ + ≤ + .
A B C k + 2 A + B + C − (k + 2)
Solution. Let us denote
L
a b c
x= , y= , z= .
.M
b c a
We need to show that
D
X x 1 4
≤ +P
x2 + kx + 1 k+2
A
x + x y + 2k − 2
P
PI
for all positive real numbers x, y, z satisfying x yz = 1. Write this inequality as follows:
M
X 1 x

2 4
− 2 ≥ −P ,
k + 2 x + kx + 1 k+2 x + x y + 2k − 2
P
LY
2 yz(x − 1)2
P
X (x − 1)2
≥P ,
O
x2 + kx + 1 x + x y + 2k − 2
P
.M
X (x − 1)2 [−x + y + z + x( y + z) − yz − 2]
≥ 0.
x2 + kx + 1
Since
W
−x + y + z + x( y + z) − yz − 2 = (x + 1)( y + z) − (x + yz + 2)
W
= (x + 1)( y + z) − (x + 1)( yz + 1) = −(x + 1)( y − 1)(z − 1),

W

X x2 − 1
−(x − 1)( y − 1)(z − 1) ≥ 0,
x2 + kx + 1
or E ≥ 0, where
X
E = −(x − 1)( y − 1)(z − 1) (x 2 − 1)( y 2 + k y + 1)(z 2 + kz + 1).
We have X
(x 2 − 1)( y 2 + k y + 1)(z 2 + kz + 1) =
X X X X
= k(2 − k) xy− x + x2 y2 − x2
= −k(2 − k)(x − 1)( y − 1)(z − 1) − (x 2 − 1)( y 2 − 1)(z 2 − 1)

= −(x − 1)( y − 1)(z − 1)[(x + 1)( y + 1)(z + 1) + k(2 − k)],
and hence
E = (x − 1)2 ( y − 1)2 (z − 1)2 [(x + 1)( y + 1)(z + 1) + k(2 − k)] ≥ 0,
because
p p p
(x + 1)( y + 1)(z + 1) + k(2 − k) ≥ (2 x)(2 y)(2 z) + k(2 − k) = (2 + k)(4 − k) ≥ 0.
L
D
.M
A
1 1 1 1 1 1
b2 + bc + c 2
+ 2
c + ca + a 2
+ 2
a + ab + b 2
≥ 2 PI
+ 2 + 2
2a + bc 2b + ca 2c + a b
.
M
LY
Solution. Write the inequality as follows:

X 1 1

O
− ≥ 0,
b2 + bc + c 2 2a2 + bc
.M
X (a2 − b2 ) + (a2 − c 2 )
≥ 0,
(b2 + bc + c 2 )(2a2 + bc)
W
X a2 − b2 X b2 − a2
+ ≥ 0,
(b2 + bc + c 2 )(2a2 + bc) (c 2 + ca + a2 )(2b2 + ca)
W
X c(a2 − b2 )(a − b)
W
(a2 + b2 + c 2 − a b − bc − ca) ≥ 0.
(b2 + bc + c 2 )(c 2 + ca + a2 )(2a2 + bc)(2b2 + ca)
Clearly, the last form is obvious. The equality holds for a = b = c.
1 1 1 1 1 1
+ + ≥ 2 + 2 + 2 .
2a b + 1 2bc + 1 2ca + 1 a +2 b +2 c +2
1 3 X 1 1
X
− ≥ − ,
2a b + 1 2 a2 + 2 2
X 1 X a2 3
+ ≥ .
2a b + 1 2(a + 2) 2
2
Let us denote
q = a b + bc + ca, q ≤ 3.
L
X 1 9 9
.M
≥P =
2a b + 1 (2a b + 1) 2q + 3
D
and P 2
a2 a 9
A
X
≥P = .
2(a + 2)
2 2(a + 2) 2(15 − 2q)
2

PI
M
9 9 3
+ ≥ .
LY
2q + 3 2(15 − 2q) 2
This inequality is true because it reduces to the obvious inequality

O
(3 − q)(9 − 2q) ≥ 0.
.M

W
W

W
1 1 1 1 1 1
+ + ≥ 2 + 2 + 2 .
a b + 2 bc + 2 ca + 2 a +2 b +2 c +2
First Solution (by Nguyen Van Quy). Rewrite the inequality as follows:
X 2 1 1

− − ≥ 0,
a b + 2 a2 + 2 b2 + 2
X a(a − b) b(b − a)

+ ≥ 0,
(a b + 2)(a2 + 2) (a b + 2)(b2 + 2)
X (2 − a b)(a − b)2 (c 2 + 2)
≥ 0.
ab + 2
Without loss of generality, assume that a ≥ b ≥ c ≥ 0. Then,
a(b + c) (a + b + c)2
bc ≤ ac ≤ ≤ =2
2 8
and
X (2 − a b)(a − b)2 (c 2 + 2) (2 − a b)(a − b)2 (c 2 + 2) (2 − ac)(a − c)2 (b2 + 2)
≥ +
ab + 2 ab + 2 ac + 2
(2 − a b)(a − b) (c + 2) (2 − ac)(a − b)2 (c 2 + 2)
2 2
L
≥ +
ab + 2 ab + 2
.M
(4 − a b − ac)(a − b)2 (c 2 + 2)
= ≥ 0.
ab + 2
D
The equality holds for a = b = c = 4/3, and also for a = 2 and b = c = 1 (or any cyclic
A
permutation).
PI
M
1 3 X 1 1
X
− ≥ − ,
bc + 2 2 a2 + 2 2
LY
X 1 X a2 3
+
O
≥ .
bc + 2 2(a2 + 2) 2
.M
Assume that a ≥ b ≥ c and denote

b+c 4
s= , p = bc, 0≤s≤ , 0 ≤ p ≤ s2 .
2 3
W

W
b2 c2 (b + c)2 s2
+ =
W
≥ .
2(b2 + 2) 2(c 2 + 2) 2(b2 + 2) + 2(c 2 + 2) + 4 2s2 − p + 2
In addition,
1 1 a(b + c) + 4 2as + 4
+ = = 2 .
ca + 2 a b + 2 (a b + 2)(ac + 2) a p + 4as + 4
Therefore, it suffices to show that E(a, b, c) ≥ 0, where
1 2(as + 2) a2 s2 3
E(a, b, c) = + 2 + + − .
p + 2 a p + 4as + 4 2(a2 + 2) 2s2 − p + 2 2
We will prove that

E(a, b, c) ≥ E(a, s, s) ≥ 0.
We have
1 1 1 1

E(a, b, c) − E(a, s, s) = − + 2(as + 2) 2 −
p + 2 s2 + 2 a p + 4as + 4 a2 s2 + 4as + 4
1 1

+ s2 − 2
2s − p + 2 s + 2
2
2
s −p 2a2 (as + 2)(s2 − p)
= +
(p + 2)(s2 + 2) (a2 p + 4as + 4)(a2 s2 + 4as + 4)
s2 (s2 − p)
− 2 .
(s + 2)(2s2 − p + 2)
L
Since s2 − p ≥ 0, it remains to show that
.M
1 2a2 (as + 2) s2
+ ≥ ,
D
(p + 2)(s2 + 2) (a2 p + 4as + 4)(a2 s2 + 4as + 4) (s2 + 2)(2s2 − p + 2)
A
2a2 (as + 2)
PI
p(s2 + 1) − 2
≥ .
M
(a2 p + 4as + 4)(a2 s2 + 4as + 4) (p + 2)(s2 + 2)(2s2 − p + 2)
LY
Since
a2 p + 4as + 4 ≤ a2 s2 + 4as + 4 = (as + 2)2
O
and
.M
2s2 − p + 2 ≥ s2 + 2,
it is enough to prove that

W
2a2 p(s2 + 1) − 2
≥ .
W
(as + 2)3 (p + 2)(s2 + 2)2

W
In addition, since
as + 2 = (4 − 2s)s + 2 ≤ 4
and
p(s2 + 1) − 2 2(s2 + 2) 2(s2 + 2)
= s2 + 1 − ≤ s2 + 1 − 2 = s2 − 1,
p+2 p+2 s +2
a2 s2 − 1
≥ 2 ,
32 (s + 2)2
(2 − s)2 (2 + s2 )2 ≥ 8(s2 − 1).
4
Indeed, for the non-trivial case 1 < s ≤ , we have
3
4 2 4

2 2 2 2
(2 − s) (2 + s ) − 8(s − 1) ≥ 2 − (2 + s2 )2 − 8(s2 − 1) = (s4 − 14s2 + 22)
3 9
2
4 2 2
4 16 88
= (7 − s ) − 27 ≥ 7− − 27 = > 0.
9 9 9 729
To end the proof, we need to show that E(a, s, s) ≥ 0. Notice that E(a, s, s) can be find
from E(a, b, c) by replacing p with s2 . We get
a2 s2
L
1 2 3
E(a, s, s) = + + + −
s2 + 2 as + 2 2(a2 + 2) s2 + 2 2
.M
(s − 1)2 (3s − 4)2
= ≥ 0.
2(s2 + 2)(1 + 2s − s2 )(2s2 − 8s + 9)
D
A
PI
M
LY
a b + bc + ca (a − b)2 (b − c)2 (c − a)2

(a) + ≤ 1;
a2 + b2 + c 2 (a2 + b2 )(b2 + c 2 )(c 2 + a2 )
O
a b + bc + ca (a − b)2 (b − c)2 (c − a)2

+
.M
(b) ≤ 1.
a2 + b2 + c 2 (a2 − a b + b2 )(b2 − bc + c 2 )(c 2 − ca + a2 )

W
Solution. (a) First Solution. Consider the non-trivial case where a, b, c are distinct and
W
write the inequality as follows:

W
(a − b)2 (b − c)2 (c − a)2 (a − b)2 + (b − c)2 + (c − a)2

≤ ,
(a2 + b2 )(b2 + c 2 )(c 2 + a2 ) 2(a2 + b2 + c 2 )
(a2 + b2 ) + (b2 + c 2 ) + (c 2 + a2 ) (a − b)2 + (b − c)2 + (c − a)2

≤ ,
(a2 + b2 )(b2 + c 2 )(c 2 + a2 ) (a − b)2 (b − c)2 (c − a)2
X 1 X 1
≤ .
(b2 + c 2 )(c 2 + a2 ) (b − c)2 (c − a)2
Since
a2 + b2 ≥ (a − b)2 , b2 + c 2 ≥ (b − c)2 , c 2 + a2 ≥ (c − a)2 ,
the conclusion follows. The equality holds for a = b = c.
Second Solution. Assume that a ≥ b ≥ c. We have
a b + bc + ca (a − b)2 (b − c)2 (c − a)2 a b + bc + ca (a − b)2 (a − c)2

+ ≤ +
a2 + b2 + c 2 (a2 + b2 )(b2 + c 2 )(c 2 + a2 ) a2 + b2 + c 2 (a2 + b2 )(a2 + c 2 )
2a b + c 2 (a − b)2 a2
≤ 2 +
a + b2 + c 2 a2 (a2 + b2 + c 2 )
2a b + c 2 + (a − b)2
= = 1.
a2 + b2 + c 2
(b) Consider the non-trivial case where a, b, c are distinct and write the inequality
as follows:
L
(a − b)2 (b − c)2 (c − a)2 (a − b)2 + (b − c)2 + (c − a)2
.M
≤ ,
(a2 − a b + b2 )(b2 − bc + c 2 )(c 2 − ca + a2 ) 2(a2 + b2 + c 2 )
D
2(a2 + b2 + c 2 ) (a − b)2 + (b − c)2 + (c − a)2
≤ ,
A
(a2 − a b + b2 )(b2 − bc + c 2 )(c 2 − ca + a2 ) (a − b)2 (b − c)2 (c − a)2
X 1
≥
PI
2(a2 + b2 + c 2 )
.
(a − b)2 (a − c)2 (a2 − a b + b2 )(b2 − bc + c 2 )(c 2 − ca + a2 )
M
Assume that a = min{a, b, c} and use the substitution
LY
b = a + x, c = a + y, x, y ≥ 0.
O
The inequality can be written as

.M
1 1 1
+ + 2 ≥ 2 f (a),
x2 y2 x 2 (x − y)2 y (x − y)2
W
where
W
3a2 + 2(x + y)a + x 2 + y 2

f (a) = .
(a2 + x a + x 2 )(a2 + y a + y 2 )[a2 + (x + y)a + x 2 − x y + y 2 ]
W
We will show that

1 1 1
+ + 2 ≥ 2 f (0) ≥ 2 f (a).
x2 y2 x 2 (x − y)2 y (x − y)2
We have
1 1 1 2(x 2 + y 2 − x y) 2(x 2 + y 2 )
+ + − 2 f (0) = −
x 2 y 2 x 2 (x − y)2 y 2 (x − y)2 x 2 y 2 (x − y)2 x 2 y 2 (x 2 − x y + y 2 )
2
= ≥ 0.
(x − y) (x − x y + y 2 )
2 2
Also, since
(a2 + x a + x 2 )(a2 + y a + y 2 ) ≥ (x 2 + y 2 )a2 + x y(x + y)a + x 2 y 2
and
a2 + (x + y)a + x 2 − x y + y 2 ≥ x 2 − x y + y 2 ,
we get f (a) ≤ g(a), where
3a2 + 2(x + y)a + x 2 + y 2

g(a) = .
[(x 2 + y 2 )a2 + x y(x + y)a + x 2 y 2 ](x 2 − x y + y 2 )
L
Therefore,
.M
x2 + y2
f (0) − f (a) ≥ − g(a)
D
x 2 y 2 (x 2 − x y + y 2 )
(x 4 − x 2 y 2 + y 4 )a2 + x y(x + y)(x − y)2 a
A
= 2 2 2 ≥ 0.
x y (x − x y + y 2 )[(x 2 + y 2 )a2 + x y(x + y)a + x 2 y 2 ]
Thus, the proof is completed. The equality holds for a = b = c.

PI
M
LY
O
.M
1 1 1 45
+ 2 + 2 ≥ .
a2 +b 2 b +c 2 c +a 2 8(a + b + c ) + 2(a b + bc + ca)
2 2 2
W

W
First Solution (by Nguyen Van Quy). Multiplying by a2 +b2 +c 2 , the inequality becomes
W
X a2 45(a2 + b2 + c 2 )
+ 3 ≥ .
b2 + c 2 8(a2 + b2 + c 2 ) + 2(a b + bc + ca)
Applying the Cauchy-Schwarz inequality, we have

P 2 2
X a2 a (a2 + b2 + c 2 )2
≥ = .
b2 + c 2 a2 (b2 + c 2 ) 2(a2 b2 + b2 c 2 + c 2 a2 )
P
(a2 + b2 + c 2 )2 45(a2 + b2 + c 2 )
+ 3 ≥ ,
2(a2 b2 + b2 c 2 + c 2 a2 ) 8(a2 + b2 + c 2 ) + 2(a b + bc + ca)
(a2 + b2 + c 2 )2 45(a2 + b2 + c 2 )
− 3 ≥ − 9,
a2 b2 + b2 c 2 + c 2 a2 4(a2 + b2 + c 2 ) + a b + bc + ca
a4 + b4 + c 4 − a2 b2 − b2 c 2 − c 2 a2 9(a2 + b2 + c 2 − a b − bc − ca)
≥ .
a2 b2 + b2 c 2 + c 2 a2 4(a2 + b2 + c 2 ) + a b + bc + ca
By Schur’s inequality of degree four, we have
a4 + b4 + c 4 − a2 b2 − b2 c 2 − c 2 a2 ≥ (a2 + b2 + c 2 − a b − bc − ca)(a b + bc + ca) ≥ 0.
L
.M
[4(a2 + b2 + c 2 ) + a b + bc + ca](a b + bc + ca) ≥ 9(a2 b2 + b2 c 2 + c 2 a2 ).
D
Since
(a b + bc + ca)2 ≥ a2 b2 + b2 c 2 + c 2 a2 ,
A
this inequality is true if PI
4(a2 + b2 + c 2 )(a b + bc + ca) ≥ 8(a2 b2 + b2 c 2 + c 2 a2 ),
M
LY
a b(a − b)2 + bc(b − c)2 + ca(c − a)2 + a bc(a + b + c) ≥ 0.

O
.M
tion).
Second Solution. Write the inequality as f6 (a, b, c) ≥ 0, where
W
X 2 Y
f6 (a, b, c) = 8(a2 + b2 + c 2 ) + 2(a b + bc + ca) (a + b2 )(a2 + c 2 ) − 45 (b2 + c 2 ).

W
Clearly, f6 (a, b, c) has the same highest coefficient A as

W
Y Y
f (a, b, c) = −45 (b2 + c 2 ) = −45 (p2 − 2q − a2 ),
where p = a + b + c and q = a b + bc + ca; that is,
A = 45.
Since A > 0, we will apply the highest coefficient cancellation method. We have
f6 (a, 1, 1) = 4a(2a + 5)(a2 + 1)(a − 1)2 ,
f6 (0, b, c) = (b − c)2 [8(b4 + c 4 ) + 18bc(b2 + c 2 ) + 15b2 c 2 ].

Since
f6 (1, 1, 1) = f6 (0, 1, 1) = 0,
define the homogeneous function
P(a, b, c) = a bc + B(a + b + c)3 + C(a + b + c)(a b + bc + ca)
such that P(1, 1, 1) = P(0, 1, 1) = 0; that is,

1 4
P(a, b, c) = a bc + (a + b + c)3 − (a + b + c)(a b + bc + ca).
9 9
We will show that the following sharper inequality holds
L
.M
f6 (a, b, c) ≥ 45P 2 (a, b, c).
Let us denote
D
g6 (a, b, c) = f6 (a, b, c) − 45P 2 (a, b, c).
A
Clearly, g6 (a, b, c) has the highest coefficient A = 0. By P 3.76-(a) in Volume 1, it suffices
PI
to prove that g6 (a, 1, 1) ≥ 0 and g6 (0, b, c) ≥ 0 for all a, b, c ≥ 0. We have
a(a − 1)2
M
P(a, 1, 1) = ,
9
LY
hence
O
a(a − 1)2 (67a3 + 190a2 + 67a + 180)

g6 (a, 1, 1) = f6 (a, 1, 1) − 45P 2 (a, 1, 1) = ≥ 0.
9
.M
Also, we have
(b + c)(b − c)2
P(0, b, c) = ,
W
9
hence
W
g6 (0, b, c) = f6 (0, b, c) − 45P 2 (0, b, c)

W
(b − c)2 [67(b4 + c 4 ) + 162bc(b2 + c 2 ) + 145b2 c 2 ]

= ≥ 0.
9
P 1.158. If a, b, c are real numbers, no two of which are zero, then
a2 − 7bc b2 − 7ca c 2 − 7a b 9(a b + bc + ca)

+ 2 + 2 + ≥ 0.
b2 + c 2 a + b2 a + b2 a2 + b2 + c 2
Solution. Let
X
f8 (a, b, c) =(a2 + b2 + c 2 ) (a2 − 7bc)(a2 + b2 )(a2 + c 2 )
Y
+ 9(a b + bc + ca) (b2 + c 2 )
is a symmetric homogeneous polynomial of degree eight. Always, f8 (a, b, c) can be

written in the form
f8 (a, b, c) = A(p, q)r 2 + B(p, q)r + C(p, q),
L
.M
where the highest polynomial A(p, q) has the form αp2 + βq. Since
X
f8 (a, b, c) =(p2 − 2q) (a2 − 7bc)(p2 − 2q − c 2 )(p2 − 2q − b2 )
D
Y
+ 9q (p2 − 2q − a2 ),
A
f8 (a, b, c) has the same highest polynomial as
X
PI
M
g8 (a, b, c) =(p2 − 2q) (a2 − 7bc)b2 c 2 + 9q(−a2 b2 c 2 )
X
=(p2 − 2q) 3r 2 − 7
LY
b3 c 3 − 9qr 2 ;
that is,
O
A(p, q) = (p2 − 2q)(3 − 21) − 9q = −9(p2 − 3q).

.M
Since A(p, q) ≤ 0 for all real a, b, c, by Lemma below, it suffices to prove that f8 (a, 1, 1) ≥
0 for all a, b, c ≥ 0. We have
W
f8 (a, 1, 1) = (a2 + 1)(a − 1)2 (a + 2)2 (a2 − 2a + 3) ≥ 0.

W
The equality holds for a = b = c, and also for −a/2 = b = c (or any cyclic permutation).
W
Lemma. Let
and let f8 (a, b, c) be a symmetric homogeneous polynomial of degree eight written in the
form
where A(p, q) ≤ 0 for all real a, b, c. The inequality f8 (a, b, c) ≥ 0 holds for all real
numbers a, b, c if and only if f8 (a, 1, 1) ≥ 0 for all real a.
Proof. For fixed p and q,
h8 (r) = A(p, q)r 2 + B(p, q)r + C(p, q)

is a concave quadratic function of r. Therefore, h8 (r) is minimal when r is minimal or

maximal; this is, according to P 2.53 in Volume 1, when f8 (a, 1, 1) ≥ 0 and f8 (a, 0, 0) ≥ 0
for all real a. Notice that the condition " f8 (a, 0, 0) ≥ 0 for all real a" is not necessary
because it follows from the condition " f8 (a, 1, 1) ≥ 0 for all real a" as follows:
f8 (a, 0, 0) = lim f8 (a, t, t) = lim t 8 f8 (a/t, 1, 1) ≥ 0.
t→0 t→0
Notice that A(p, q) is called the highest polynomial of f8 (a, b, c).

Remark. This Lemma can be extended as follow.
• The inequality f8 (a, b, c) ≥ 0 holds for all real numbers a, b, c satisfying A(p, q) ≤ 0
if and only if f8 (a, 1, 1) ≥ 0 for all real a such that A(a + 2, 2a + 1) ≤ 0.
L
.M
P 1.159. If a, b, c are real numbers such that a bc 6= 0, then
D
(b + c)2 (c + a)2 (a + b)2 10(a + b + c)2
A
+ + ≥ 2 + .
a2 b2 c2 3(a2 + b2 + c 2 )
PI
(Vasile Cîrtoaje and Michael Rozenberg, 2014)
M
Solution. Let
LY

O
X
f8 (a, b, c) = 3(a2 + b2 + c 2 ) b2 c 2 (b + c)2 − 2a2 b2 c 2 − 10a2 b2 c 2 (a + b + c)2 .
.M
From
X X X
b2 c 2 (b + c)2 − 2a2 b2 c 2 = b2 c 2 (p − a)2 − 2r 2 = p2 b2 c 2 − 2pqr + r 2 ,
W
it follows that f8 (a, b, c) has the same highest polynomial as

W
3(a2 + b2 + c 2 )r 2 − 10r 2 (a + b + c)2 ;

W
that is,
A(p, q) = 3(p2 − 2q) − 10p2 = −7p2 − 6q.
There are two cases to consider.
Case 1: A(p, q) ≤ 0. According to Remark from the preceding P 1.158, it suffices to show
that f8 (a, 1, 1) ≥ 0 for all real a such that A(a + 2, 2a + 1) ≤ 0. Indeed, we have
f8 (a, 1, 1) = 3(a2 + 2)[4 + 2a2 (a + 1)2 − 2a2 ] − 10a2 (a + 2)2
= 2(3a6 + 6a5 + a4 − 8a3 − 14a2 + 12)
= 2(a − 1)2 (3a4 + 12a3 + 22a2 + 24a + 12)
= 2(a − 1)2 [3(a + 1)4 + (2a + 3)2 ] ≥ 0.
Case 2: A(p, q) > 0. We will show that there exist two real numbers B and C such that
the following sharper inequality holds:
f8 (a, b, c) ≥ A(p, q)P 2 (a, b, c),
where
P(a, b, c) = r + Bp3 + C pq.
Let us denote
g8 (a, b, c) = f8 (a, b, c) − A(p, q)P 2 (a, b, c).
We see that the highest polynomial of g8 (a, b, c) is zero. Thus, according to Remark
from P 1.158, it suffices to prove that g(a) ≥ 0 for all real a, where g(a) = g8 (a, 1, 1).
L
We have
.M
g(a) = f8 (a, 1, 1) − A(a + 2, 2a + 1)P 2 (a, 1, 1),
where
D
f8 (a, 1, 1) = 2(3a6 + 6a5 + a4 − 8a3 − 14a2 + 12),
A
A(a + 2, 2a + 1) = −7(a + 2)2 − 6(2a + 1) = −7a2 − 40a − 34,
PI
P(a, 1, 1) = a + B(a + 2)3 + C(a + 2)(2a + 1).
M
Since
g(−2) = 72 − 18(−2)2 = 0,
LY
a necessary condition to have g(a) ≥ 0 in the vicinity of −2 is g 0 (−2) = 0. This condition

involves C = −1. We can check that g 00 (−2) = 0 for this value of C. Thus, a necessary
O
condition to have g(a) ≥ 0 in the vicinity of −2 is g 000 (−2) = 0 is necessary. This

condition involves B = 2/3. For these values of B and C, we have
.M
2
2

2 3
g(a) = f8 (a, 1, 1) + (7a + 40a + 34) a + (a + 2) − (a + 2)(2a + 1)
3
W
2 4 4 3 2
= (a + 2) (14a + 52a + 117a + 154a + 113).
W
9
Since
W
14a4 + 52a3 + 117a2 + 154a + 113 = (a2 + 1)2 + 13a2 (a + 2)2 + 7(9a2 + 22a + 16) > 0,
the proof is completed. The equality holds for a = b = c.
a2 − 4bc b2 − 4ca c 2 − 4a b 9(a b + bc + ca) 9

+ 2 + 2 + ≥ .
b2 + c 2 a + b2 a + b2 a2 + b2 + c 2 2
Solution. Let
X
f8 (a, b, c) =2(a2 + b2 + c 2 ) (a2 − 4bc)(a2 + b2 )(a2 + c 2 )
Y
+ 9(2a b + 2bc + 2ca − a2 − b2 − c 2 ) (b2 + c 2 )
is a symmetric homogeneous polynomial of degree eight. Always, f8 (a, b, c) can be

written in the form
L
.M
where A(p, q) = αp2 + βq is called the highest polynomial of f8 (a, b, c). Since
X
f8 (a, b, c) =2(p2 − 2q) (a2 − 4bc)(p2 − 2q − c 2 )(p2 − 2q − b2 )
D
Y
A
+ 9(4q − p2 (p2 − 2q − a2 ),
f8 (a, b, c) has the same highest polynomial as

PI
M
X
g8 (a, b, c) =2(p2 − 2q) (a2 − 4bc)b2 c 2 + 9(4q − p2 )(−a2 b2 c 2 )
LY
X
=2(p2 − 2q) 3r 2 − 4 b3 c 3 − 9(4q − p2 )r 2 ;
O
that is,
A(p, q) = 2(p2 − 2q)(3 − 12) − 9(4q − p2 ) = −9p2 .
.M
Since A(p, q) ≤ 0 for all a, b, c ≥ 0, by Lemma below, it suffices to prove that f8 (a, 1, 1) ≥
0 and f8 (0, b, c) ≥ 0 for all a, b, c ≥ 0. We have
W
f8 (a, 1, 1) = 2a(a + 4)(a2 + 1)(a − 1)4 ≥ 0

W
and
f8 (0, b, c) = (b2 + c 2 )(2E − 9F ),
W
where
E = −4b3 c 3 + (b2 + c 2 )(b4 + c 4 ), F = b2 c 2 (b − c)2 .
Since
E ≥ −4b3 c 3 + 2bc(b4 + c 4 ) = 2bc(b2 − c 2 )2 ,
we have
2E − 9F ≥ 4bc(b2 − c 2 )2 − 9b2 c 2 (b − c)2 = bc(b − c)2 [4(b + c)2 − 9bc] ≥ 0.
Thus, the proof is completed. The equality holds for a = b = c, and also for a = 0 and
Lemma. Let
and let f8 (a, b, c) be a symmetric homogeneous polynomial of degree eight written in the
form
where A(p, q) ≤ 0 for all a, b, c ≥ 0. The inequality f8 (a, b, c) ≥ 0 holds for all nonnegative
real numbers a, b, c if and only if f8 (a, 1, 1) ≥ 0 and f8 (0, b, c) ≥ 0 for all a, b, c ≥ 0.
Proof. For fixed p and q,
h8 (r) = A(p, q)r 2 + B(p, q)r + C(p, q)
L
.M
is a concave quadratic function of r. Therefore, h8 (r) is minimal when r is minimal or
maximal. This is, according to P 3.57 in Volume 1, when b = c or a = 0. Thus, the
conclusion follows. Notice that A(p, q) is called the highest polynomial of f8 (a, b, c).
D
Remark. This Lemma can be extended as follow.
A
• The inequality f8 (a, b, c) ≥ 0 holds for all a, b, c ≥ 0 satisfying A(p, q) ≤ 0 if and only
PI
if f8 (a, 1, 1) ≥ 0 and f8 (0, b, c) ≥ 0 for all a, b, c ≥ 0 such that A(a + 2, 2a + 1) ≤ 0 and
M
A(b + c, bc) ≤ 0.
LY
O
.M
a2 + b2 + c 2 9(a − b)2 (b − c)2 (c − a)2

≥1+ .
a b + bc + ca (a + b)2 (b + c)2 (c + a)2
W
Solution. Consider the non-trivial case where a, b, c are distinct and a = min{a, b, c}.
W
Write the inequality as follows:

W
(a − b)2 + (b − c)2 + (c − a)2 9(a − b)2 (b − c)2 (c − a)2

≥ ,
2(a b + bc + ca) (a + b)2 (b + c)2 (c + a)2
(a − b)2 + (b − c)2 + (c − a)2 18(a b + bc + ca)

≥ ,
(a − b) (b − c) (c − a)
2 2 2 (a + b)2 (b + c)2 (c + a)2
X 1 18(a b + bc + ca)
≥ .
(b − a) (c − a)
2 2 (a + b)2 (a + c)2 (b + c)2
Since
X 1 1 1 1 2(b2 + c 2 − bc)
≥ + + =
(b − a)2 (c − a)2 b2 c 2 b2 (b − c)2 c 2 (b − c)2 b2 c 2 (b − c)2
and
a b + bc + ca a b + bc + ca 1
≤ ≤ ,
(a + b) (a + c) (b + c)
2 2 2 (a b + bc + ca) (b + c)
2 2 bc(b + c)2
b2 + c 2 − bc 9
≥ .
b c (b − c)
2 2 2 bc(b + c)2
Write this inequality as follows:
(b + c)2 − 3bc 9(b + c)2 − 36bc

≥ ,
bc (b + c)2
L
(b + c)2 36bc
.M
− 12 + ≥ 0,
bc (b + c)2
(b + c)4 − 12bc(b + c)2 + 36b2 c 2 ≥ 0,
D
[(b + c)2 − 6bc]2 ≥ 0.
A
PI
b/c + c/b = 4 (or any cyclic permutation).
M
LY
O
a2 + b2 + c 2 p (a − b)2 (b − c)2 (c − a)2

.M
≥ 1 + (1 + 2)2 2 .
a b + bc + ca (a + b2 )(b2 + c 2 )(c 2 + a2 )
W

p
Solution. Consider the non-trivial case where a, b, c are distinct and denote k = 1+ 2.
W
Write the inequality as follows:

W
(a − b)2 + (b − c)2 + (c − a)2 k2 (a − b)2 (b − c)2 (c − a)2

≥ 2 ,
2(a b + bc + ca) (a + b2 )(b2 + c 2 )(c 2 + a2 )
(a − b)2 + (b − c)2 + (c − a)2 2k2 (a b + bc + ca)

≥ ,
(a − b)2 (b − c)2 (c − a)2 (a2 + b2 )(b2 + c 2 )(c 2 + a2 )
X 1 2k2 (a b + bc + ca)
≥ .
(b − a)2 (c − a)2 (a2 + b2 )(b2 + c 2 )(c 2 + a2 )
Assume that a = min{a, b, c}, and use the substitution
b = a + x, c = a + y, x, y ≥ 0.
The inequality becomes
1 1 1
+ + 2 ≥ 2k2 f (a),
x2 y2 x 2 (x − y)2 y (x − y)2
where
3a2 + 2(x + y)a + x y

f (a) = .
(2a2 + 2x a + x 2 )(2a2 + 2 y a + y 2 )[2a2 + 2(x + y)a + x 2 + y 2 ]
We will show that

1 1 1
L
+ + 2 ≥ 2k2 f (0) ≥ 2k2 f (a).
x2 y2 x 2 (x − y)2 y (x − y)2
.M
We have
D
1 1 1 2 2(x 2 + y 2 − x y) 2k2 x y
+ + (0) =
A
− 2k f −
x 2 y 2 x 2 (x − y)2 y 2 (x − y)2 x 2 y 2 (x − y)2 x 2 y 2 (x 2 + y 2 )
PI p
2[x 2 + y 2 − (2 + 2 )x y]2
= 2 2 ≥ 0.
x y (x − y)2 (x 2 − x y + y 2 )
M
Also, since
LY
(2a2 + 2x a + x 2 )(2a2 + 2 y a + y 2 ) ≥ 2(x 2 + y 2 )a2 + 2x y(x + y)a + x 2 y 2

O
and
.M
2a2 + 2(x + y)a + x 2 + y 2 ≥ x 2 + y 2 ,

we get f (a) ≤ g(a), where
W
3a2 + 2(x + y)a + x y

g(a) = .
W
[2(x 2 + y 2 )a2 + 2x y(x + y)a + x 2 y 2 ](x 2 + y 2 )

W
Therefore,
1
f (0) − f (a) ≥ − g(a)
x y(x 2 + y 2 )
(2x 2 + 2 y 2 − 3x y)a2
= ≥ 0.
x y(x 2 + y 2 )[2(x 2 + y 2 )a2 + 2x y(x + y)a + x 2 y 2 ]
pcompleted. The equality holds for a = b = c, and also for a = 0 and

Thus, the proof is
b/c + c/b = 2 + 2 (or any cyclic permutation).
2 2 2 5 5 5
+ + ≥ + + .
a+b b+c c+a 3a + b + c 3b + c + a 3c + a + b
Solution. Write the inequality as follows:

X 2 5

− ≥ 0,
b + c 3a + b + c
X 2a − b − c
≥ 0,
(b + c)(3a + b + c)
L
X a−b X a−c
.M
+ ≥ 0,
(b + c)(3a + b + c) (b + c)(3a + b + c)
X a−b X b−a
+
D
≥ 0,
(b + c)(3a + b + c) (c + a)(3b + c + a)
A
X (a − b)2 (a + b − c)
PI ,
(b + c)(c + a)(3a + b + c)(3b + c + a)
X
(b − c)2 Sa ≥ 0,
M
where
LY
Sa = (b + c − a)(b + c)(3a + b + c).

Assume that a ≥ b ≥ c. Since Sc > 0, it suffices to show that
O
(b − c)2 Sa + (a − c)2 S b ≥ 0.
.M
Since S b ≥ 0 and (a − c)2 ≥ (b − c)2 , we have

W
(b − c)2 Sa + (a − c)2 S b ≥ (b − c)2 Sa + (b − c)2 S b = (b − c)2 (Sa + S b ).

W
Thus, it is enough to prove that Sa + S b ≥ 0, which is equivalent to

W
(c + a − b)(c + a)(3b + c + a) ≥ (b + c − a)(b + c)(3a + b + c).
Consider the nontrivial case where b + c − a > 0. Since c + a − b ≥ b + c − a, we only

need to show that
(c + a)(3b + c + a) ≥ (b + c)(3a + b + c).
Indeed,
(c + a)(3b + c + a) − (b + c)(3a + b + c) = (a − b)(a + b − c) ≥ 0.

8a2 + 3bc 8b2 + 3ca 8c 2 + 3a b
(a) + + ≥ 11;
b2 + bc + c 2 c 2 + ca + a2 a2 + a b + b2
8a2 − 5bc 8b2 − 5ca 8c 2 − 5a b

(b) + + ≥ 9.
b2 − bc + c 2 c 2 − ca + a2 a2 − a b + b2
Solution. Consider the more general inequality
a2 + mbc b2 + mca c 2 + ma b 3(m + 1)

+ + ≥ .
L
b + k bc + c
2 2 c + kca + a
2 2 a + ka b + b
2 2 k+2
.M
Let p = a + b + c and q = a b + bc + ca. Write the inequality in the form f6 (a, b, c) ≥ 0,
where
D
X
f6 (a, b, c) = (k + 2) (a2 + mbc)(a2 + ka b + b2 )(a2 + kac + c 2 )
A
Y
−3(m + 1) (b2 + k bc + c 2 ).
From
PI
M
X
f6 (a, b, c) = (k + 2) (a2 + mbc)(ka b − c 2 + p2 − 2q)(kac − b2 + p2 − 2q)
LY
Y
−3(m + 1) (k bc − a2 + p2 − 2q).
O
(k + 2)P2 (a, b, c) − 3(m + 1)P3 (a, b, c),

.M
where X
P2 (a, b, c) = (a2 + mbc)(ka b − c 2 )(kac − b2 ),
W
Y
P3 (a, b, c) = (k bc − a2 ).
W

W
A = (k + 2)P2 (1, 1, 1) − 3(m + 1)P3 (1, 1, 1)

= 3(k + 2)(m + 1)(k − 1)2 − 3(m + 1)(k − 1)3 = 9(m + 1)(k − 1)2 .
Also, we have
f6 (a, 1, 1) = (k + 2)(a2 + ka + 1)(a − 1)2 [a2 + (k + 2)a + 1 + 2k − 2m].
(a) For our particular case m = 3/8 and k = 1, we have A = 0. Therefore, according
to P 2.75 in Volume 1, it suffices to prove that f6 (a, 1, 1) ≥ 0 for all real a. Indeed,
3 2

f6 (a, 1, 1) = 3(a2 + a + 1)(a − 1)2 a + ≥ 0.
2
Thus, the proof is completed. The equality holds for a = b = c, and also for −2a/3 =
(b) For m = −5/8 and k = −1, we have A = 27/2 and
1 2
f6 (a, 1, 1) = (a − a + 1)(a − 1)2 (2a + 1)2 .
4
Since A > 0, we will use the highest coefficient cancellation method. Define the homoge-
neous polynomial
P(a, b, c) = r + Bp3 + C pq,
L
where B and C are real constants. Since the desired inequality becomes an equality
for a = b = c = 1, and also for a = −1 and b = c = 2, determine B and C such that
.M
P(1, 1, 1) = P(−1, 2, 2) = 0. We find
D
4 −5
B= , C= ,
27 9
A
when
P(a, 1, 1) =
2
(a − 1)2 (2a + 1),
PI
P(a, 0, 0) = Ba3 =
4 3
a .
M
27 27
We will show that
LY
27 2
f6 (a, b, c) ≥ P (a, b, c).
2
O
Let us denote
27 2
g6 (a, b, c) = f6 (a, b, c) − P (a, b, c).
.M
2
Since g6 (a, b, c) has the highest coefficient A = 0, it suffices to prove that g6 (a, 1, 1) ≥ 0
for all real a (see P 2.75 in Volume 1). Indeed,
W
27 2 1
W
g6 (a, 1, 1) = f6 (a, 1, 1) − P (a, 1, 1) = (a − 1)2 (2a + 1)2 (19a2 − 11a + 19) ≥ 0.

2 108
W
Thus, the proof is completed. The equality holds for a = b = c, and also for −2a = b = c
4a2 + bc 4b2 + ca 4c 2 + a b
+ + ≥ 1.
4b2 + 7bc + 4c 2 4c 2 + 7ca + 4a2 4a2 + 7a b + 4b2
Solution. Write the inequality as f6 (a, b, c) ≥ 0, where

X Y
f6 (a, b, c) = (4a2 + bc)(4a2 + 7a b + 4b2 )(4a2 + 7ac + 4c 2 ) − (4b2 + 7bc + 4c 2 ).
Let
From
X
f6 (a, b, c) = (4a2 + bc)(7a b − 4c 2 + 4p2 − 8q)(7ac − 4b2 + 4p2 − 8q)
Y
− (7bc − 4a2 + 4p2 − 8q),
L
.M
P2 (a, b, c) − P3 (a, b, c),
D
A
where X
P2 (a, b, c) = (4a2 + bc)(7a b − 4c 2 )(7ac − 4b2 ),
PI
Y
P3 (a, b, c) = (7bc − 4a2 ).
M
LY
A = P2 (1, 1, 1) − P3 (1, 1, 1) = 135 − 27 = 108.

O
Since A > 0, we will apply the highest coefficient cancellation method. Define the homo-
.M
geneous polynomial
P(a, b, c) = r + Bp3 + C pq,
W
where B and C are real constants. We will show that there are two real numbers B and
C such that the following sharper inequality holds
W
f6 (a, b, c) ≥ 108P 2 (a, b, c).

W
Let us denote
g6 (a, b, c) = f6 (a, b, c) − 108P 2 (a, b, c).
Clearly, g6 (a, b, c) has the highest coefficient A1 = 0. Then, by P 2.75 in Volume 1, it
suffices to prove that g6 (a, 1, 1) ≥ 0 for all real a.
We have
g6 (a, 1, 1) = f6 (a, 1, 1) − 108P 2 (a, 1, 1),
where
f6 (a, 1, 1) = 4(4a2 + 7a + 4)(a − 1)2 (4a2 + 15a + 16),
P(a, 1, 1) = a + B(a + 2)3 + C(a + 2)(2a + 1).
Let us denote g(a) = f6 (a, 1, 1). Since g(−2) = 0, we can have g(a) ≥ 0 in the vicinity
of a = −2 only if g 0 (−2) = 0, which involves C = −5/9. On the other hand, from
g(1) = 0, we get B = 4/27. For these values of B and C, we get
2(a − 1)2 (2a + 1)

P(a, 1, 1) = ,
27
4
g6 (a, 1, 1) = (a − 1)2 (a + 2)2 (416a2 + 728a + 431) ≥ 0.
27
The proof is completed. The equality holds for a = b = c, and for a = 0 and b + c = 0
L
D
.M
P 1.166. If a, b, c are real numbers, no two of which are equal, then
A
1 1 1 27
+ + ≥ PI .
(a − b)2 (b − c)2 (c − a)2 4(a + b + c − a b − bc − ca)
2 2 2
M
LY
1 1 1 27

(a − b)2 + (b − c)2 + (a − c)2

+ +
O
≥ ,
(a − b)2 (b − c)2 (a − c)2 2
.M
(a − b)2 (b − c)2 (a − c)2 (a − c)2

27
+ +1 + +1 ≥ ,
(a − c)2 (a − c)2 (a − b)2 (b − c)2 2
W
1 1 27

(x 2 + y 2 + 1) + + 1 ≥ ,
x2 y2 2
W
where
W
a−b b−c
x= , y= , x + y = 1.
a−c a−c
We have
1 1 27 (x + 1)2 (x − 2)2 (2x − 1)2

2 2
(x + y + 1) + + 1 − = ≥ 0.
x2 y2 2 2x 2 (1 − x)2
The proof is completed. The equality holds for 2a = b + c (or any cyclic permutation).
Second Solution. Assume that a > b > c. We have
1 1 2 8 8
+ ≥ ≥ = .
(a − b)2 (b − c)2 (a − b)(b − c) [(a − b) + (b − c)] 2 (a − c)2

9 27
≥ ,
(a − c)2 4(a + b + c − a b − bc − ca)
2 2 2
(a − 2b + c)2 ≥ 0.
Third Solution. Write the inequality as f6 (a, b, c) ≥ 0, where

X
f6 (a, b, c) = 4(a2 + b2 + c 2 − a b − bc − ca) (a − b)2 (a − c)2 −27(a − b)2 (b − c)2 (c − a)2 .
L
.M
−27(a − b)2 (b − c)2 (c − a)2 ;
that is,
D
A = −27(−27) = 729.
A
Since A > 0, we will use the highest coefficient cancellation method. Define the homoge-
neous polynomial PI
1

M
P(a, b, c) = a bc + B(a + b + c)3 − 3B + (a + b + c)(a b + bc + ca),
9
LY
which satisfies the property P(1, 1, 1) = 0. We will show that there is a real value of B
such that the following sharper inequality holds
O
f6 (a, b, c) ≥ 729P 2 (a, b, c).

.M
Let us denote
g6 (a, b, c) = f6 (a, b, c) − 729P 2 (a, b, c).
Clearly, g6 (a, b, c) has the highest coefficient A1 = 0. Then, by P 2.75 in Volume 1, it
W
suffices to prove that g6 (a, 1, 1) ≥ 0 for all real a.

W
We have
f6 (a, 1, 1) = 4(a − 1)6
W
and
1
P(a, 1, 1) = (a − 1)2 [9B(a + 2) + 2],
9
hence
g6 (a, 1, 1) = f6 (a, 1, 1)−729P 2 (a, 1, 1) = (27B +2)(a −1)4 (a +2)[(2−27B)a −54B −8].
Choosing B = −2/27, we get g6 (a, 1, 1) = 0 for all real a.
Remark. The inequality is equivalent to
(a − 2b + c)2 (b − 2c + a)2 (c − 2a + b)2 ≥ 0.

1 1 1 14
+ 2 + 2 ≥ .
a2 − ab + b 2 b − bc + c 2 c − ca + a 2 3(a + b2 + c 2 )
2
(Vasile Cîrtoaje and BJSL, 2014)

X
f6 (a, b, c) = 3(a2 + b2 + c 2 ) (a2 − a b + b2 )(a2 − ac + c 2 )
−14(a2 − a b + b2 )(b2 − bc + c 2 )(c 2 − ca + a2 ).

L
.M
−14(a2 − a b + b2 )(b2 − bc + c 2 )(c 2 − ca + a2 ),
hence as
D
f (a, b, c) = −14(−c 2 − a b)(−a2 − bc)(−b2 − ca);
A
that is, according to Remark 2 from the proof of P 2.75 in Volume 1,
PI
A = f (1, 1, 1) = −14(−2)3 = 112.
M
Since A > 0, we apply the highest coefficient cancellation method. Define the homoge-
LY
neous polynomial
P(a, b, c) = a bc + B(a + b + c)3 + C(a + b + c)(a b + bc + ca).

O
We will show that there are two real numbers B and C such that the following sharper
.M
inequality holds
f6 (a, b, c) ≥ 112P 2 (a, b, c).
W
Let us denote
g6 (a, b, c) = f6 (a, b, c) − 112P 2 (a, b, c).
W
Clearly, g6 (a, b, c) has the highest coefficient A1 = 0. By P 2.75 in Volume 1, it suffices

to prove that g6 (a, 1, 1) ≥ 0 for all real a.
W
We have
g6 (a, 1, 1) = f6 (a, 1, 1) − 112P 2 (a, 1, 1),
where
f6 (a, 1, 1) = (a2 − a + 1)(3a4 − 3a3 + a2 + 8a + 4),
P(a, 1, 1) = 1 + B(a + 2)3 + C(a + 2)(2a + 1).
Let us denote g(a) = g6 (a, 1, 1). Since g(−2) = 0, we can have g(a) ≥ 0 in the vicinity
of a = −2 only if g 0 (−2) = 0, which involves C = −4/7. In addition, setting B = 9/56,
we get
1
P(a, 1, 1) = (9a3 − 10a2 + 4a + 8),
56
3 6
g6 (a, 1, 1) = (a + 4a5 + 8a4 + 16a3 + 20a2 + 16a + 16)
28
3(a + 2)2 (a2 + 2)2
= ≥ 0.
28
The proof is completed. The equality holds for a = 0 and b + c = 0 (or any cyclic
permutation).
P 1.168. Let a, b, c be real numbers such that a b + bc + ca ≥ 0 and no two of which are
L
zero. Prove that
.M
a b c 3
(a) + + ≥ ;
b+c c+a a+b 2
D
(b) i f a b ≤ 0, then
A
a b c
+ + ≥ 2. PI
b+c c+a a+b
M
LY
Solution. Let as show first that b + c 6= 0, c + a 6= 0 and a + b 6= 0. Indeed, if b + c = 0,

then a b + bc + ca ≥ 0 yields b = c = 0, which is not possible.
O
(a) Write the inequality as follows

.M
X a 9
+1 ≥ ,
b+c 2
W
X X 1

(b + c) ≥ 9,
W
b+c
Xa + b a + c
W
+ − 2 ≥ 0,
a+c a+b
X (b − c)2
≥ 0,
(a + b)(a + c)
X (b − c)2
≥ 0.
a2 + (a b + bc + ca)
Clearly, the last inequality is true. The equality holds for a = b = c 6= 0.
(b) From a b + bc + ca ≥ 0, it follows that if one of a, b, c is zero, then the others
are the same sign. In this case, the desired inequality is trivial. So, due to symmetry
and homogeneity, it suffices to consider that a < 0 < b ≤ c.
First Solution. We will show that
F (a, b, c) > F (0, b, c) ≥ 2,
where
a b c
F (a, b, c) = + + .
b+c c+a a+b
We have
b c
F (0, b, c) = + ≥2
c b
and
L
1

b c
F (a, b, c) − F (0, b, c) = a
.M
− − .
b + c c(c + a) b(a + b)
Since a < 0, we need to show that
D
b c 1
A
+ > .
c(c + a) b(a + b) b+c
From a b + bc + ca ≥ 0, we get
PI
M
−ca −a b
c+a≥ > 0, a+b≥ > 0,
LY
b c
O
hence
b b c c
> 2, > 2.
.M
c(c + a) c b(a + b) b
W
b c 1
+ 2≥ .
2 b+c
W
c b
Indeed, by virtue of the AM-GM inequality, we have

W
b c 1 2 1
+ 2− ≥ p − p > 0.
c 2 b b+c bc 2 bc
This completes the proof. The equality holds for a = 0 and b = c, or b = 0 and a = c.
Second Solution. Since b + c > 0 and
(b + c)(a + b) = b2 + (a b + bc + ca) > 0, (b + c)(c + a) = c 2 + (a b + bc + ca) > 0,
we get a + b > 0 and c + a > 0. By virtue of the Cauchy-Schwarz inequality and AM-GM
inequality, we have
a b c a (b + c)2
+ + ≥ +
b+c c+a a+b b + c b(c + a) + c(a + b)
a (b + c)2
> +
2a + b + c (b + c)2
+ a(b + c)
2
4a 2(b + c)
> + = 2.
2a + b + c 2a + b + c
L
.M
a b c a b + bc + ca
+ +
D
≥ .
7a + b + c 7b + c + a 7c + a + b (a + b + c)2
A
First Solution. Write the inequality as follows:

PI
M
X 2a a(b + c)

− ≥ 0,
LY
7a + b + c (a + b + c)2
X a[(a − b) + (a − c)](a − b − c)
O
≥ 0,
7a + b + c
.M
X a(a − b)(a − b − c) X a(a − c)(a − b − c)

+ ≥ 0,
7a + b + c 7a + b + c
W
X a(a − b)(a − b − c) X b(b − a)(b − c − a)

+ ≥ 0,
7a + b + c 7b + c + a
W
a(a − b − c) b(b − c − a)
X
(a − b) − ≥ 0,
7a + b + c 7b + c + a
W
X
(a − b)2 (a2 + b2 − c 2 + 14a b)(a + b + 7c) ≥ 0.
Since
a2 + b2 − c 2 + 14a b ≥ (a + b)2 − c 2 = (a + b + c)(a + b − c),
X
(a − b)2 (a + b − c)(a + b + 7c) ≥ 0.
Assume that a ≥ b ≥ c. It suffices to show that
(a − c)2 (a − b + c)(a + 7b + c) + (b − c)2 (−a + b + c)(7a + b + c) ≥ 0.

For the nontrivial case b > 0, we have
a2 a
(a − c)2 ≥ 2
(b − c)2 ≥ (b − c)2 .
b b
a(a − b + c)(a + 7b + c) + b(−a + b + c)(7a + b + c) ≥ 0.
Since
a(a + 7b + c) ≥ b(7a + b + c),
L
we have
.M
a(a − b + c)(a + 7b + c) + b(−a + b + c)(7a + b + c) ≥
D
≥ b(a − b + c)(7a + b + c) + b(−a + b + c)(7a + b + c)
A
= 2bc(7a + b + c) ≥ 0.
PI
This completes the proof. The equality holds for a = b = c, and also for a = 0 and b = c
M
Second Solution. Assume that a ≤ b ≤ c, a + b + c = 3 and use the substitution

LY
2a + 1 2b + 1 2c + 1
x= y= , z=
O
, ,
3 3 3
.M
where 1/3 ≤ x ≤ y ≤ z, x + y + z = 3. We have b + c ≥ 2, y + z ≥ 2, x ≤ 1. The

inequality becomes
W
a b c 9 − a2 − b2 − c 2
+ + ≥ ,
2a + 1 2b + 1 2c + 1
W
1 1 1

W
9(x 2 + y 2 + z 2 ) ≥ 4 + + + 17.
x y z
Assume that x ≤ y ≤ z and show that
E(x, y, z) ≥ E(x, t, t) ≥ 0,
where
t = ( y + z)/2 = (3 − x)/2
and
1 1 1

2 2 2
E(x, y, z) = 9(x + y + z ) − 4 + + − 17.
x y z
We have
1 1 2

2 2 2
E(x, y, z) − E(x, t, t) = 9( y + z − 2t ) − 4 + −
y z t
( y − z) [9 yz( y + z) − 8]
2
= ≥ 0,
2 yz( y + z)
since
9 yz = (2b + 1)(2c + 1) ≥ 2(b + c) + 1 ≥ 5, y + z ≥ 2.
L
Also,
.M
4 8 (x − 1)2 (3x − 1)(8 − 3x)
E(x, t, t) = 9x 2 + 2t 2 − 15 − − = ≥ 0.
x t 2x(3 − x)
D
Third Solution. Write the inequality as f5 (a, b, c) ≥ 0, where f5 (a, b, c) is a symmetric
A
homogeneous inequality of degree five. According to P 3.68-(a) in Volume 1, it suffices
PI
to prove the inequality for a = 0 and for b = c = 1. For a = 0, the inequality is
equivalent to
M
(b − c)2 (b2 + c 2 + 11bc) ≥ 0,
LY
while, for b = c = 1, the inequality is equivalent to

O
a(a − 1)2 (a + 14) ≥ 0.

.M
W

W
a2 b2 c2 1
+ + ≥ .
W
4a + 5bc 4b + 5ca 4c + 5a b
2 2 2 3

X Y
f6 (a, b, c) = 3 a2 (4b2 + 5ca)(4c 2 + 5a b) − (4a2 + 5bc)
X X
= −45a2 b2 c 2 − 25a bc a3 + 40 a3 b3 .
Since f6 (a, b, c has the highest coefficient
A = −45 − 75 + 120 = 0,
according to 2.67-(b), it suffices to prove the original inequality b = c = 1 and 0 ≤ a ≤ 2,

and for a = b + c.
Case 1: b = c = 1, 0 ≤ a ≤ 2. The original inequality reduces to
(2 − a)(a − 1)2 ≥ 0,
which is true.
Case 2: a = b + c. Using the Cauchy-Schwarz inequality
b2 c2 (b + c)2
+ ≥ ,
4b2 + 5ca 4c 2 + 5a b 4(b2 + c 2 ) + 5a(b + c)
L
.M
a2 (b + c)2 1
+
D
≥ ,
4a2 + 5bc 4(b2 + c 2 ) + 5a(b + c) 3
A
PI
(b − c)2 (3b2 + 3c 2 − 4bc) ≥ 0.
M
The equality holds for an equilateral triangle, and for a degenerate triangle with a/2 =
LY

O
.M

1 1 1 3
+ 2 + 2
W
≥ .
7a2 +b +c
2 2 7b + c + a
2 2 7c + a + b
2 2 (a + b + c)2
W

W

where
X Y
f6 (a, b, c) = p2 (7b2 + c 2 + a2 )(7c 2 + a2 + b2 ) − 3 (7a2 + b2 + c 2 )
X Y
= p2 (6b2 + p2 − 2q)(6c 2 + p2 − 2q) − 3 (6a2 + p2 − 2q).
Since f6 (a, b, c has the highest coefficient
A = −3(63 ) < 0,
according to 2.67-(b), it suffices to prove the original inequality b = c = 1 and 0 ≤ a ≤ 2,

and for a = b + c.
a(8 − a)(a − 1)2 ≥ 0,
which is true.
Case 2: a = b + c. Write the inequality as
1 1 1 3
+ 2 + 2 ≥ ,
4(b2 + c ) + 7bc 4b + c + bc 4c + b + bc
2 2 2 2(b + c)2
1 5x + 2 3
+ ≥ ,
4x + 7 4x + 5x + 10 2(x + 2)
L
2
.M
where x = b/c + c/b, x ≥ 2. This inequality is equivalent to the obvious inequality
16x 2 + 5x − 38 ≥ 0.
D
The equality holds for an equilateral triangle, and for a degenerate triangle with a = 0
A
PI
M
LY
P 1.172. Let a, b, c be the lengths of the sides of a triangle. If k > −2, then
X a(b + c) + (k + 1)bc 3(k + 3)
O
≤ .
b2 + k bc + c2 k+2
.M

W
where Y
f6 (a, b, c) = 3(k + 3) (b2 + k bc + c 2 )
W
X
−(k + 2) [a(b + c) + (k + 1)bc](c 2 + kca + a2 )(a2 + ka b + b2 )
W
Y
= 3(k + 3) (p2 − 2q + k bc − a2 )
X
−(k + 2) (q + k bc)(p2 − 2q + kca − b2 )(p2 − 2q + ka b − c 2 ).
Since f6 (a, b, c has the same highest coefficient A as f (a, b, c), where
Y X
f (a, b, c) = 3(k + 3) (k bc − a2 ) − k(k + 2) bc(kca − b2 )(ka b − c 2 )
X X
= 3(k + 3)[(k3 − 1)a2 b2 c 2 − k2 a bc a3 + k a3 b3 ]
X X
−k(k + 2)(3k2 a2 b2 c 2 − 2ka bc a3 + a3 b3 ),
we get
A = 3(k + 3)(k3 − 1 − 3k2 + 3k) − k(k + 2)(3k2 − 6k + 3) = −9(k − 1)2 ≤ 0.
According to 2.67-(b), it suffices to prove the original inequality b = c = 1 and 0 ≤ a ≤

2, and for a = b + c.
(2 − a)(a − 1)2 ≥ 0,
which is true.
L
Case 2: a = b + c. Write the inequality as follows
.M
X a(b + c) + (k + 1)bc
3
−1 ≤ ,
b + k bc + c k+2
D
2 2
A
X a b + bc + ca − b2 − c 2 3
≤ ,
3bc
b2
bc − c 2
+ k bc + c2 PI
bc − b2
k+2
3
+ +
M
≤ .
b + k bc + c
2 2 b + (k + 2)(bc + c ) c + (k + 2)(bc + b )
2 2 2 2 k+2
LY
Since
3bc 3
≤ ,
b2 + k bc + c 2 k+2
O

.M
bc − c 2 bc − b2
+ ≤ 0.
b2 + (k + 2)(bc + c 2 ) c 2 + (k + 2)(bc + b2 )
W
This reduces to the obvious inequality

W
(b − c)2 (b2 + bc + c 2 ) ≥ 0.
W
P 1.173. Let a, b, c be the lengths of the sides of a triangle. If k > −2, then
X 2a2 + (4k + 9)bc 3(4k + 11)
≤ .
b2 + k bc + c2 k+2

where Y
f6 (a, b, c) = 3(4k + 11) (b2 + k bc + c 2 )
X
−(k + 2) [2a2 + (4k + 9)bc](c 2 + kca + a2 )(a2 + ka b + b2 )
Y
= 3(4k + 11) (p2 − 2q + k bc − a2 )
X
−(k + 2) [2a2 + (4k + 9)bc](p2 − 2q + kca − b2 )(p2 − 2q + ka b − c 2 ).
Since f6 (a, b, c has the same highest coefficient A as f (a, b, c), where
L
Y
f (a, b, c) = 3(4k + 11) (k bc − a2 )
.M
X
−(k + 2) [2a2 + (4k + 9)bc](kca − b2 )(ka b − c 2 )
D
X X
= 3(4k + 11)[(k3 − 1)a2 b2 c 2 − k2 a bc a3 + k a3 b3 ]
A
X X
PI
−(k + 2)[3(4k3 + 9k2 + 2)a2 b2 c 2 − 6k(k + 3)a bc a3 + 9 a3 b3 ],
we get
M
A = 3(4k + 11)(k3 − 1 − 3k2 + 3k) − (k + 2)[3(4k3 + 9k2 + 2) − 18k(k + 3) + 27]
LY
= −9(4k + 11)(k − 1)2 ≤ 0.

O
According to 2.67-(b), it suffices to prove the original inequality b = c = 1 and 0 ≤ a ≤

.M
2, and for a = b + c.
W
(2 − a)(a − 1)2 ≥ 0,
W
which is true.
W
Case 2: a = b + c. Write the inequality as follows

X 2a2 + (4k + 9)bc
3(2k + 7)
−2 ≤ ,
b + k bc + c
2 2 k+2
X 2(a2 − b2 − c 2 ) + (2k + 9)bc 3(2k + 7)
≤ ,
b2 + k bc + c2 k+2
(2k + 13)bc

c b
+ (2k + 5)(b + c) +
b2 + k bc + c 2 b2 + (k + 2)(bc + c 2 ) c 2 + (k + 2)(bc + b2 )
3(2k + 7)
≤ .
k+2
Using the substitution x = b/c + c/b, x ≥ 2, the inequality can be written as

2k + 13 (2k + 5)(x + 2)(x + 2k + 3) 3(2k + 7)
+ ≤ ,
x +k (k + 2)x + (k + 2)(k + 3)x + 2k + 6k + 5
2 2 k+2
(x − 2)[4(k + 2)Ax 2 + 2(k + 2)B x + C] ≥ 0,

where
A = k + 4, B = 2k2 + 13k + 22, C = 8k3 + 51k2 + 98k + 65.
The inequality is true since A > 0,
B = 2(k + 2)2 + 5(k + 2) + 4 > 0,
L
C = 8(k + 2)3 + 2k2 + (k + 1)2 > 0.
.M
D
A
P 1.174. If a ≥ b ≥ c ≥ d such that a bcd = 1, then
PI
M
1 1 1 3
+ + ≥ p .
LY
1+a 1+ b 1+c 3
1 + a bc
O
Solution. We can get this inequality by summing the inequalities below

.M
1 1 2
+ ≥ p ,
1+a 1+ b 1 + ab
W
1 2 3
+ p ≥ p .
1 + c 1 + ab 3
1 + a bc
W
The first inequality is true, since

W
1 1 2 1 1 1 1

+ − p = − p + − p
1 + a 1 + b 1 + ab 1 + a 1 + ab 1 + b 1 + ab
p p p
( a − b)2 ( a b − 1)
= p
(1 + a)(1 + b)(1 + a b)
p p
and
p a b ≥ a bcd = 1. To prove the second inequality, we denote x = a b andpy =
3
a bc (x ≥ y ≥ 1), which yield c = y 3 /x 2 . From a bc 2 ≥ a bcd = 1, we get a bc ≥ a b,
that is, y 3 ≥ x. Since
1 2 3 x2 2 3
+ p − p = + −
1 + c 1 + a b 1 + a bc
3
x +y
2 3 1+ x 1+ y
x2

1 1 1

= − +2 −
x +y
2 3 1+ y 1+ x 1+ y
(x − y)2 [( y − 2)x + 2 y 2 − y]
= ,
(1 + x)(1 + y)(x 2 + y 3 )
we still have to show that ( y − 2)x + 2 y 2 − y ≥ 0. This is clearly true for y ≥ 2, while
for 1 ≤ y < 2, we have
( y − 2)x + 2 y 2 − y ≥ ( y − 2) y 3 + 2 y 2 − y = y( y − 1)( y 2 − y + 1) ≥ 0.
L
.M
D
P 1.175. Let a, b, c, d be positive real numbers such that a bcd = 1. Prove that
A
X 1
1 + a b + bc + ca
≤ 1. PI
M
LY
Solution. From
1 1 1 1 1 1 p p p p
O
+ + ≥p +p +p = d( a + b + c),
a b c bc ca ab
.M
we get
p p p
a+ b+ c
a b + bc + ca ≥ p .
W
d
Therefore,
W
p
X 1 X d
≤ p p p p = 1,
1 + a b + bc + ca a+ b+ c+ d
W
which is just the required inequality. The equality occurs for a = b = c = d = 1.
1 1 1 1
+ + + ≥ 1.
(1 + a)2 (1 + b)2 (1 + c)2 (1 + d)2

First Solution. The inequality follows by summing the following inequalities (see P
1.1):
1 1 1
+ ≥ ,
(1 + a)2 (1 + b)2 1 + ab
1 1 1 ab
+ ≥ = .
(1 + c)2 (1 + d)2 1 + cd 1 + ab
The equality occurs for a = b = c = d = 1.
Second Solution. Using the substitutions a = 1/x 4 , b = 1/ y 4 , c = 1/z 4 , d = 1/t 4 ,
where x, y, z, t are positive real numbers such that x yz t = 1, the inequality becomes as
follows
L
x6 y6 z6 t6
+ + + 2 ≥ 1.
.M
2 2 2
1 1 1 1

x3 + y3 + z3 + t3 +
x y z t
D
A
( x 3 )2 ( x 3 )2
P P
X x6
2 ≥ 2 = P 6 .
PI
x + 2 x 2 + x 2 y 2z2
P P
1 1

x3 + x3 +
P
x x
M
LY
X X
2(x 3 y 3 + x 3 z 3 + x 3 t 3 + y 3 z 3 + y 3 t 3 + z 3 t 3 ) ≥ 2x yz t x2 + x 2 y 2z2.
O
This is true if
.M
X
2(x 3 y 3 + x 3 z 3 + x 3 t 3 + y 3 z 3 + y 3 t 3 + z 3 t 3 ) ≥ 3x yz t x2
and
W
X
2(x 3 y 3 + x 3 z 3 + x 3 t 3 + y 3 z 3 + y 3 t 3 + z 3 t 3 ) ≥ 3 x 2 y 2z2,
W
Write these inequalities as

W
X
x 3 ( y 3 + z 3 + t 3 − 3 yz t) ≥ 0
and X
(x 3 y 3 + y 3 z 3 + z 3 x 3 − 3x 2 y 2 z 2 ) ≥ 0,
respectively. By the AM-GM inequality, we have y 3 + z 3 + t 3 ≥ 3 yz t and x 3 y 3 + y 3 z 3 +

z 3 x 3 ≥ 3x 2 y 2 z 2 . Thus the conclusion follows.
Third Solution. Using the substitutions a = yz/x 2 , b = z t/ y 2 , c = t x/z 2 , d = x y/t 2 ,
where x, y, z, t are positive real numbers, the inequality becomes
x4 y4 z4 t4
+ + + ≥ 1.
(x 2 + yz)2 ( y 2 + z t)2 (z 2 + t x)2 (t 2 + x y)2
Using the Cauchy-Schwarz inequality two times, we deduce
x4 z4 x4 z4
+ ≥ +
(x 2 + yz)2 (z 2 + t x)2 (x 2 + y 2 )(x 2 + z 2 ) (z 2 + t 2 )(z 2 + x 2 )
x4 z4 x 2 + z2

1
= 2 + ≥ ,
x + z2 x 2 + y 2 z2 + t 2 x 2 + y 2 + z2 + t 2
and hence
x4 z4 x 2 + z2
+ ≥ .
(x 2 + yz)2 (z 2 + t x)2 x 2 + y 2 + z2 + t 2
Adding this to the similar inequality
L
.M
y4 t4 y2 + t2
+ ≥ ,
( y 2 + z t)2 (t 2 + x y)2 x 2 + y 2 + z2 + t 2
D
we get the required inequality.
A
Fourth Solution. Using the substitutions a = x/ y, b = y/z, c = z/t, d = t/x, where
PI
x, y, z, t are positive real numbers, the inequality can be written as
y2 z2 t2 x2
M
+ + + ≥ 1.
(x + y)2 ( y + z)2 (z + t)2 (t + x)2
LY
By the Cauchy-Schwarz inequality and the AM-GM inequality, we get

O
[ y( y + z)]2
P
X y2
≥P
(x + y)2 (x + y)2 ( y + z)2
.M
[(x + y)2 + ( y + z)2 + (z + t)2 + (t + x)2 ]2

= ≥ 1.
4[(x + y)2 + (z + t)2 ][( y + z)2 + (t + x)2 ]
W
Remark. The following generalization holds true (Vasile Cîrtoaje, 2005):

W
p
• Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. If k ≥ n − 1,
W
then
1 1 1 n
+ + ··· + ≥ .
(1 + ka1 ) 2 (1 + ka2 )2 (1 + kan )2 (1 + k)2
1
P 1.177. Let a, b, c, d 6= be positive real numbers such that a bcd = 1. Prove that
3
1 1 1 1
+ + + ≥ 1.
(3a − 1)2 (3b − 1)2 (3c − 1)2 (3d − 1)2
First Solution. It suffices to show that
1 a−3
≥ .
(3a − 1)2 a−3 + b−3 + c −3 + d −3
6a−2 + b−3 + c −3 + d −3 ≥ 9a−1 ,
which follows from the AM-GM inequality, as follows

9
p
6a−2 + b−3 + c −3 + d −3 ≥ 9 a−12 b−3 c −3 d −3 = 9a−1 .
L
.M
Second Solution. Let a ≤ b ≤ c ≤ d. If a < 1/3, then
D
1
A
> 1,
(3a − 1)2
PI
and the desired inequality is clearly true. Otherwise, if 1/3 < a ≤ b ≤ c ≤ d, we have
M
4a3 − (3a − 1)2 = (a − 1)2 (4a − 1) ≥ 0.
LY
Therefore, using this result and the AM-GM inequality, we get

O
v
X 1 1X 1 t
4 1
.M
≥ ≥ = 1.
(3a − 1)2 4 a 3 a b c3 d 3
3 3
W
Third Solution. We have

W
1 1 a(a − 1)2 (a + 2)(a2 + 3)

− = ≥ 0.
(3a − 1)2 (a3 + 1)2 (3a − 1)2 (a3 + 1)2
W
Therefore,
X 1 X 1
≥ ,
(3a − 1)2 (a + 1)2
3
and it suffices to prove that

X 1
≥ 1.
(a3 + 1)2
This inequality is an immediate consequence of the inequality in P 1.176.
1 1 1 1
+ + + ≥ 1.
1 + a + a2 + a3 1 + b + b2 + b3 1 + c + c 2 + c 3 1 + d + d 2 + d 3
First Solution. We get the desired inequality by summing the inequalities

1 1 1
+ ≥ ,
1+a+a +a
2 3 1+ b+ b + b
2 3 1 + (a b)3/2
1 1 1
+ ≥ .
L
1+c+c +c
2 3 1+d +d +d
2 3 1 + (cd)3/2
.M
1 1 1
+
D
≥ ,
1 + x2 + x4 + x6 1 + y2 + y4 + y6 1 + x3 y3
A
where x and y are positive real numbers. Putting p = x y and s = x 2 + x y + y 2 , this
inequality becomes PI
M
p3 (x 6 + y 6 ) + p2 (p − 1)(x 4 + y 4 ) − p2 (p2 − p + 1)(x 2 + y 2 ) − p6 − p4 + 2p3 − p2 + 1 ≥ 0,
LY
p3 (x 3 − y 3 )2 + p2 (p − 1)(x 2 − y 2 )2 − p2 (p2 − p + 1)(x − y)2 + p6 − p4 − p2 + 1 ≥ 0.

p3 s2 (x − y)2 + p2 (p − 1)(s + p)2 (x − y)2 − p2 (p2 − p + 1)(x − y)2 + p6 − p4 − p2 + 1,
O
p2 (s + 1)(ps − 1)(x − y)2 + (p2 − 1)(p4 − 1) ≥ 0.

.M
If ps − 1 ≥ 0, then this inequality is clearly true. Consider further that ps < 1. From
ps < 1 and s ≥ 3p, we get p2 < 1/3. Write the desired inequality in the form
W
(1 − p2 )(1 − p4 ) ≥ p2 (1 + s)(1 − ps)(x − y)2 .

W
Since
p(x − y)2 = p(s − 3p) < 1 − 3p2 < 1 − p2 ,
W

1 − p4 ≥ p(1 + s)(1 − ps).
Indeed,
4p(1 + s)(1 − ps) ≤ [p(1 + s) + (1 − ps)]2 = (1 + p)2 < 2(1 + p2 ) < 4(1 − p4 ).

Second Solution. Assume that a ≥ b ≥ c ≥ d, and write the inequality as
X 1
≥ 1.
(1 + a)(1 + a2 )
Since
1 1 1 1 1 1
≤ ≤ , ≤ ≤ ,
1+a 1+ b 1+c 1+a 2 1+ b 2 1 + c2
by Chebyshev’s inequality, it suffices to prove that
1 1 1 1 1 1 1 1

+ + + + + ≥ 1.
3 1+a 1+ b 1+c 1 + a2 1 + b2 1 + c 2 (1 + d)(1 + d 2 )
In addition, from the inequality in P 1.174, we have
p3
1 1 1 3 3 d
+ + ≥ p = p
1+a 1+ b 1+c 3
1 + a bc
3
d +1
L
.M
and p3
1 1 1 3 3 d2
+ + ≥ p = p .
1 + a2 1 + b2 1 + c 2 3
1 + a2 b2 c 2
3
d2 + 1
D
A
3d 1
(1 +
p
3
d)(1 +
p
3
d2 )
+ PI
(1 + d)(1 + d 2 )
≥ 1.
M
p
3
Putting x = d, this inequality becomes as follows
LY
3x 3 1
+ ≥ 1,
(1 + x)(1 + x 2 ) (1 + x 3 )(1 + x 6 )
O
3x 3 (1 − x + x 2 )(1 − x 2 + x 4 ) + 1 ≥ (1 + x 3 )(1 + x 6 ),
.M
x 3 (2 − 3x + 2x 3 − 3x 5 + 2x 6 ) ≥ 0,
x 3 (1 − x)2 (2 + x + x 3 + 2x 4 ) ≥ 0.
W
Remark. The following generalization holds true (Vasile Cîrtoaje, 2004):

W
• If a1 , a2 , . . . , an are positive real numbers such that a1 a2 · · · an = 1, then

W
1 1 1
+ + ··· + ≥ 1.
1 + a1 + · · · + a1n−1 1 + a2 + · · · + a2n−1 1 + an + · · · + ann−1
1 1 1 1
+ + + ≥ 1.
1 + a + 2a2 1 + b + 2b2 1 + c + 2c 2 1 + d + 2d 2
Solution. We will show that

1 1
≥ ,
1 + a + 2a 2 1 + a + a2k + a3k
k
where k = 5/6. Then, it suffices to show that

X 1
≥ 1,
1 + a k + a2k + a3k
which immediately follows from the inequality in P 1.178. Setting a = x 6 , x > 0, the
claimed inequality can be written as
L
1 1
.M
≥ ,
2x 12 + x + 1 1 + x + x 10 + x 15
6 5
D
x 10 + x 5 + 1 ≥ 2x 7 + x.
A
We can prove it by summing the AM-GM inequalities PI
M
x 5 + 4 ≥ 5x
LY
and
5x 10 + 4x 5 + 1 ≥ 10x 7 .
O
This completes the proof. The equality occurs for a = b = c = d = 1.

.M
Remark. The inequalities in P 1.176, P 1.178 and P 1.179 are particular cases of the
following more general inequality (Vasile Cîrtoaje, 2009):
W
• Let a1 , a2 , . . . , an (n ≥ 4) be positive real numbers such that a1 a2 · · · an = 1. If p, q, r

are nonnegative real numbers satisfying p + q + r = n − 1, then
W
i=n
W
X 1
≥ 1.
i=1
1 + pai + qai2 + r ai3
1 1 1 1 9 25
+ + + + ≥ .
a b c d a+b+c+d 4
Solution (by Vo Quoc Ba Can). Replacing a, b, c, d by a4 , b4 , c 4 , d 4 , respectively, the in-

equality becomes as follows:
1 1 1 1 9 25
+ 4+ 4+ 4+ 4 ≥ ,
a 4 b c d a +b +c +d
4 4 4 4a bcd
1 1 1 1 4 9 9
+ 4+ 4+ 4− ≥ − 4 ,
a 4 b c d a bcd 4a bcd a + b + c 4 + d 4
4
1 1 1 1 4 9(a4 + b4 + c 4 + d 4 − 4a bcd)
+ + + − ≥ .
a4 b4 c 4 d 4 a bcd 4a bcd(a4 + b4 + c 4 + d 4 )
Using the identities
L
a4 + b4 + c 4 + d 4 − 4a bcd = (a2 − b2 )2 + (c 2 − d 2 )2 + 2(a b − cd)2 ,
.M
1 1 1 1 4 (a2 − b2 )2 (c 2 − d 2 )2 2(a b − cd)2
+ + + − = + + 2 2 2 2 ,
D
a4 b4 c 4 d 4 a bcd a4 b4 c4 d 4 a b c d
A
(a2 − b2 )2 (c 2 − d 2 )2 2(a b − cd)2
a4 b4
+
c4 d 4
+
a2 b2 c 2 d 2
≥
PI
9[(a2 − b2 )2 + (c 2 − d 2 )2 + 2(a b − cd)2 ]
4a bcd(a4 + b4 + c 4 + d 4 )
,
M
2 2 4cd(a + b + c + d ) 2 2 4a b(a + b + c + d )
4 4 4 4 4 4 4 4

2 2
(a − b ) − 9 + (c − d ) −9
LY
a3 b3 c3 d 3
2 4(a + b + c + d )
4 4 4 4

O
+2(a b − cd) − 9 ≥ 0.
a bcd
.M
By the AM-GM inequality, we have
a4 + b4 + c 4 + d 4 ≥ 4a bcd.
W

W
2 2 4cd(a + b + c + d ) 2 2 4a b(a + b + c + d )
4 4 4 4 4 4 4 4

2 2
(a − b ) − 9 + (c − d ) − 9 ≥ 0.
W
a3 b3 c3 d 3
Without loss of generality, assume that a ≥ c ≥ d ≥ b. Since
(a2 − b2 )2 ≥ (c 2 − d 2 )2
and
4cd(a4 + b4 + c 4 + d 4 ) 4(a4 + b4 + c 4 + d 4 ) 4(a4 + 3b4 )
≥ ≥ > 9,
a3 b3 a3 b a3 b
4cd(a4 + b4 + c 4 + d 4 ) 4a b(a4 + b4 + c 4 + d 4 )

−9 + − 9 ≥ 0,
a3 b3 c3 d 3

cd ab
2(a4 + b4 + c 4 + d 4 ) 3 3
+ 3 3 ≥ 9.
a b c d
Indeed, by the AM-GM inequality,
2

4 4 4 4 cd ab
2(a + b + c + d ) 3 3 + 3 3 ≥ 8a bcd = 16 > 9.
a b c d a bcd
L
P 1.181. If a, b, c, d are real numbers such that a + b + c + d = 0, then
.M
(a − 1)2 (b − 1)2 (c − 1)2 (d − 1)2
+ + 2 + ≤ 4.
D
3a2 + 1 3b2 + 1 3c + 1 3d 2 + 1
A
Solution. Since
4−
3(a − 1)2
3a2 + 1
=
(3a + 1)2
3a2 + 1
PI
,
M
LY
X (3a + 1)2
≥ 4.
3a2 + 1
O
On the other hand, since

.M
4a2 = 3a2 + (b + c + d)2 ≤ 3a2 + 3(b2 + c 2 + d 2 ) = 3(a2 + b2 + c 2 + d 2 ),

9 2 9(a2 + b2 + c 2 + d 2 ) + 4
3a2 + 1 ≤ (a + b2 + c 2 + d 2 ) + 1 = ,
W
4 4
we have
W
4 (3a + 1)2
P
X (3a + 1)2
≥ = 4.
3a2 + 1 9(a2 + b2 + c 2 + d 2 ) + 4
W
The equality holds for a = b = c = d = 0, and also for a = 1 and b = c = d = −1/3 (or
Remark. The following generalization is also true.
• If a1 , a2 , . . . , an are real numbers such that a1 + a2 + · · · + an = 0, then
(a1 − 1)2 (a2 − 1)2 (an − 1)2
+ + ··· + ≤ n,
(n − 1)a12 + 1 (n − 1)a22 + 1 (n − 1)an2 + 1
with equality for a1 = a2 = · · · = an = 0, and also for a1 = 1 and a2 = a3 = · · · = an =
−1/(n − 1 (or any cyclic permutation).
P 1.182. If a, b, c, d ≥ −5 such that a + b + c + d = 4, then
1−a 1− b 1−c 1−d

+ + + ≥ 0.
(1 + a)2 (1 + b)2 (1 + c)2 (1 + d)2
Solution. Assume that a ≤ b ≤ c ≤ d. We show first that x ∈ [−5, −1) ∪ (−1, ∞)

involves
1− x −1
≥ ,
(1 + x)2 8
and x ∈ [−5, −1) ∪ (−1, 1/3] involves
L
1− x 3
≥ .
.M
(1 + x)2 8
Indeed, we have
D
1− x 1 (x − 3)2
+ = ≥0
A
(1 + x)2 8 8(1 + x)2
and
1− x 3 (5 + x)(1 − 3x)
− =
PI
≥ 0.
M
(1 + x)2 8 8(1 + x)2
Then, if a ≤ 1/3, then
LY
1−a 1− b 1−c 1−d 3 1 1 1

+ + + ≥ − − − = 0.
O
(1 + a)2 (1 + b)2 (1 + c)2 (1 + d)2 8 8 8 8

.M
Assume now that 1/3 ≤ a ≤ b ≤ c ≤ d. Since
1−a ≥1− b ≥1−c ≥1−d

W
and
1 1 1 1
W
≥ ≥ ≥ ,
(1 + a)2 (1 + b)2 (1 + c)2 (1 + d)2
W
by Chebyshev’s inequality, we have
1−a 1− b 1−c 1−d

+ + + ≥
(1 + a)2 (1 + b)2 (1 + c)2 (1 + d)2
1 X X 1

≥ (1 − a) = 0.
4 (1 + a)2
The equality holds for a = b = c = d = 1, and also for a = −5 and b = c = d = 3 (or
P 1.183. Let a1 , a2 , . . . , an be positive real numbers such that a1 + a2 + · · · + an = n. Prove

that X 1 1
≤ .
(n + 1)a1 + a2 + · · · + an
2 2 2 2
First Solution. By the Cauchy-Schwarz inequality, we have

X n2 X (a1 + a2 + · · · + an )2
=
(n + 1)a12 + a22 + · · · + an2 2a12 + (a12 + a22 ) + · · · + (a12 + an2 )
a22 an2

X 1
L
≤ + + ··· + 2
2 a12 + a22 a1 + an2
.M
n n(n − 1) n2
= + = ,
2 2 2
D
from which the conclusion follows. The equality holds for a1 = a2 = · · · = an = 1.
A
X a12 + a22 + · · · + an2
PI
a12 + a22 + · · · + an2
M
≤ .
(n + 1)a12 + a22 + · · · + an2 2
LY
Since
a12 + a22 + · · · + an2 na12
=1−
O
,
(n + 1)a12 + a22 + · · · + an2 (n + 1)a12 + a22 + · · · + an2
.M

X a12 a12 + a22 + · · · + an2
+ ≥ 1.
W
(n + 1)a12 + a22 + · · · + an2 2n

W

W
X a12 (a1 + a2 + · · · + an )2
≥
(n + 1)a12 + a22 + · · · + an2 [(n + 1)a12 + a22 + · · · + an2 ]
P
n
= .
2(a1 + a2 + · · · + an2 )
2 2
n a12 + a22 + · · · + an2

+ ≥ 2,
a12 + a22 + · · · + an2 n
which follows immediately from the AM-GM inequality.

P 1.184. Let a1 , a2 , . . . , an be real numbers such that a1 + a2 + · · · + an = 0. Prove that
(a1 + 1)2 (a2 + 1)2 (an + 1)2 n

+ + ··· + ≥ .
a12 + n − 1 a22 + n − 1 an + n − 1
2 n−1
Solution. Without loss of generality, assume that an2 = max{a12 , a22 , · · · , an2 }. Since
(an + 1)2 n (n − 1 − an )2
= − ,
an2 + n − 1 n − 1 (n − 1)(an2 + n − 1)
L
.M
n−1
X (ai + 1)2 (n − 1 − an )2
D
≥ .
i=1
ai2 + n − 1 (n − 1)(an2 + n − 1)
A
From the Cauchy-Schwarz inequality
n−1 n−1 n−1
PI 2
X (a + 1)2
M
X X
i
(ai2 + n − 1) ≥ (ai + 1) ,
a 2
+ n−1
LY
i=1 i=1 i i=1
we get
O
n−1
X (ai + 1)2 (n − 1 − an )2
≥ .
ai2 + n − 1 n−1 2
i=1 ai + (n − 1)
P
2
.M
i=1

W
n−1
X
ai2 + (n − 1)2 ≤ (n − 1)(an2 + n − 1),
W
i=1
−a1
W
which is clearly true. The proof is completed. The equality holds for = a2 = a3 =
n−1
· · · = an (or any cyclic permutation).
P 1.185. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. Prove that
1 1 1
+ + ··· + ≥ 1.
1 + (n − 1)a1 1 + (n − 1)a2 1 + (n − 1)an

First Solution. Let k = (n − 1)/n. We can get the required inequality by summing the
inequalities below for i = 1, 2, · · · , n:
−k
1 ai
≥ −k .
1 + (n − 1)ai −k
a1 + a2 + · · · + an−k
a1−k + · · · + ai−1
−k −k
+ ai+1 + · · · + an−k ≥ (n − 1)ai1−k ,
which follows from the AM-GM inequality. The equality holds for a1 = a2 = · · · = an =
L
1.
.M
Second Solution. Using the substitutions ai = 1/x i for all i, the inequality becomes
x1 x2 xn
+ + ··· + ≥ 1,
D
x1 + n − 1 x2 + n − 1 xn + n − 1
A
where x 1 , x 2 , · · · , x n are positive real numbers such that x 1 x 2 · · · x n = 1. By the Cauchy-
Schwarz inequality, we have PI
Pp 2
M
X xi ( x1)
≥P .
xi + n − 1 (x 1 + n − 1)
LY
Thus, we still have to prove that

O
Xp X
( x 1 )2 ≥ x 1 + n(n − 1),
.M
which reduces to X p n(n − 1)

xi x j ≥ .
W
1≤i< j≤n
2
W
Since x 1 x 2 · · · x n = 1, this inequality follows from the AM-GM inequality.

Third Solution. For the sake of contradiction, assume that the required inequality is not
W
true. Then, it suffices to show that the hypothesis a1 a2 · · · an = 1 does not hold. More
precisely, we will prove that
1 1 1
+ + ··· + <1
1 + (n − 1)a1 1 + (n − 1)a2 1 + (n − 1)an
1
involves a1 a2 · · · an > 1. Let x i = , 0 < x i < 1, for i = 1, 2, · · · , n. Since
1 + (n − 1)ai
1 − xi
ai = for all i, we need to show that
(n − 1)x i
x1 + x2 + · · · + x n < 1
implies
(1 − x 1 )(1 − x 2 ) · · · (1 − x n ) > (n − 1)n x 1 x 2 · · · x n .
Using the AM-GM inequality, we have
X vY
u
1 − xi > x k ≥ (n − 1) n−1
t xk.
k6=i k6=i
Multiplying the inequalities

vY
u
1 − x i > (n − 1) n−1 xk.
L
t
k6=i
.M
for i = 1, 2, · · · , n, the conclusion follows.
D
Remark. The inequality in P 1.185 is a particular case of the following more general
A
results (Vasile Cîrtoaje, 2005):
PI
• Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. If 0 < k ≤ n − 1
and p ≥ n1/k − 1, then
M
1 1 1 n
+ + ··· + ≥ .
LY
(1 + pa1 ) k (1 + pa2 ) k (1 + pan ) k (1 + p)k
1
O
• Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. If k ≥ and

n 1/k n−1
.M
0<p≤ − 1, then
n−1
1 1 1 n
W
+ + ··· + ≤ .
(1 + pa1 ) k (1 + pa2 ) k (1 + pan ) k (1 + p)k
W
W
1 1 1
+ + ··· + ≥ 1.
1 − a1 + na12 1 − a2 + na22 1 − an + nan2
Solution. First, we show that
1 1
≥ ,
1 − x + nx 2 1 + (n − 1)x k
1
where x > 0 and k = 2 + . Write the inequality as
n−1
(n − 1)x k + x ≥ nx 2 .
We can get this inequality using the AM-GM inequality as follows

n
p
(n − 1)x k + x ≥ n x (n−1)k x = nx 2 .

1 1 1
+ + ··· + ≥ 1,
1 + (n − 1)a1k 1 + (n − 1)a2k 2 1 + (n − 1)ank
L
.M
which follows immediately from the inequality in the preceding P 1.185. The equality
holds for a1 = a2 = · · · = an = 1.
D
Remark 1. Similarly, we can prove the following more general statement.
A
• Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. If p and q are
p
real numbers such that p + q = n − 1 and n − 1 ≤ q ≤ ( n + 1)2 , then
PI
1 1 1
M
+ + ··· + ≥ 1.
1+ pa1 + qa12 1+ pa2 + qa22 1 + pan + qan2
LY
Remark 2. We can extend the inequality in Remark 1 as follows (Vasile Cîrtoaje, 2009).
O
• Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. If p and q are

.M
p
real numbers such that p + q = n − 1 and 0 ≤ q ≤ ( n + 1)2 , then
1 1 1
+ + ··· +
W
≥ 1.
1+ pa1 + qa12 1+ pa2 + qa22 1 + pan + qan2
W
W
P 1.187. Let a1 , a2 , . . . , an be positive real numbers such that

k(n − k − 1)
a1 , a2 , . . . , an ≥ , k>1
kn − k − 1
and
a1 a2 · · · an = 1.
Prove that
1 1 1 n
+ + ··· + ≤ .
a1 + k a2 + k an + k 1+k
Solution. We use the induction on n. Let

1 1 1 n
En (a1 , a2 , . . . , an ) = + + ··· + − .
a1 + k a2 + k an + k 1 + k
For n = 2, we have
p p
(1 − k)( a1 − a2 )2
E2 (a1 , a2 ) = ≤ 0.
(1 + k)(a1 + k)(a2 + k)
Assume that the inequality is true for n−1 numbers (n ≥ 3), and prove that En (a1 , a2 , . . . , an ) ≥
0 for a1 a2 · · · an = 1 and a1 , a2 , . . . , an ≥ pn , where
L
k(n − k − 1)
.M
pn = .
kn − k − 1
Due to symmetry, we may assume that a1 ≥ 1 and a2 ≤ 1. There are two cases to
D
consider.
A
Case 1: a1 a2 ≤ k2 . Since a1 a2 ≥ a2 and pn−1 < pn , from a1 , a2 , . . . , an ≥ pn it follows
that PI
a1 a2 , a3 , · · · , an > pn−1 .
M
Then, by the inductive hypothesis, we have En−1 (a1 a2 , a2 , . . . , an ) ≤ 0, and it suffices to
LY
show that
En (a1 , a2 , . . . , an ) ≤ En−1 (a1 a2 , a2 , . . . , an ).
O
This is equivalent to
.M
1 1 1 1
+ − − ≤ 0,
a1 + k a2 + k a1 a2 + k 1 + k
W

W
(a1 − 1)(1 − a2 )(a1 a2 − k2 ) ≤ 0.

W
Case 2: a1 a2 ≥ k2 . Since
1 1 a1 + a2 + 2k a1 + a2 + 2k 1
+ = ≤ 2 =
a1 + k a2 + k a1 a2 + k(a1 + a2 ) + k2 k + k(a1 + a2 ) + k2 k
and
1 1 n−2 kn − k − 1
+ ··· + ≤ = ,
a3 + k an + k pn + k k(k + 1)
we have
1 kn − k − 1 n
En (a1 , a2 , . . . , an ) ≤ + − = 0.
k k(k + 1) 1+k
Thus, the proof is completed. The equality holds for a1 = a2 = · · · = an = 1.
Remark. For k = n − 1, we get the inequality in P 1.185. Also, for k → ∞, we get the
known inequalities
1 1 1
1< + + ··· + < n − 1,
1 + a1 1 + a2 1 + an
which holds for all positive numbers a1 , a2 , . . . , an satisfying a1 a2 · · · an = 1.
P 1.188. Let a1 , a2 , . . . , an be positive real numbers such that
L
a1 ≥ 1 ≥ a2 ≥ · · · ≥ an , a1 a2 · · · an = 1.
.M
Prove that
1 − a1 1 − a2 1 − an
+ + ··· + ≥ 0.
D
3 + a12 3 + a22 3 + an2
A
Solution. For n = 2, we have PI
(a1 − 1)4
M
1 − a1 1 − a2
+ = ≥ 0.
3 + a12 3 + a22 (3 + a12 )(3a12 + 1)
LY
For n ≥ 3, we will use the induction method and the inequality

O
1− x 1− y 1− xy
+ ≥ ,
3+ x 2 3+ y 2 3 + x2 y2
.M
which holds for all x, y ∈ [0, 1]. Indeed, we can write this inequality as
(1 − x)(1 − y)(3 + x y)(3 − x y − x 2 y − x y 2 ) ≥ 0,

W
W
which is obviously true. Based on the inequality

1 − an−1 1 − an 1 − an−1 an
W
+ ≥ ,
3 + an−1
2 3 + an2 3 + an−1
2
an2

1 − a1 1 − an−2 1 − an−1 an
+ ··· + + ≥ 0.
3 + a12 3 + an−2
2
3 + an−1
2
an2
Since
a1 ≥ 1 ≥ a2 ≥ · · · an−2 ≥ an−1 an ,
this inequality follows from the hypothesis induction. Thus, the proof is completed. The
equality holds for a1 = a2 = · · · = an = 1.
P 1.189. If a1 , a2 , . . . , an ≥ 0, then
1 1 1 n
+ + ··· + ≥ .
1 + na1 1 + na2 1 + nan n + a1 a2 · · · an
Solution. If one of a1 , a2 , . . . , an is zero, the inequality is obvious. Consider further that
a1 , a2 , . . . , an > 0 and let
r = n a1 a2 · · · an .
p

Pp Pp
L
X 1 ( a2 a3 · · · an )2 ( a2 a3 · · · an )2
≥P =P .
.M
1 + na1 (1 + na1 )a2 a3 · · · an a2 a3 · · · an + n2 r n
D
Xp X
(n + r n )( a2 a3 · · · an )2 ≥ n a2 a3 · · · an + n3 r n .
A
Xp X
PI
( a2 a3 · · · an )2 ≥ a2 a3 · · · an + n(n − 1)r n−1 .
M
LY
X X
(n + r n )[ a2 a3 · · · an + n(n − 1)r n−1 ] ≥ n a2 a3 · · · an + n3 r n ,
O
.M
X
rn a2 a3 · · · an + n(n − 1)r 2n−1 + n2 (n − 1)r n−1 ≥ n3 r n .
W
Also, by the AM-GM inequality,

X
a2 a3 · · · an ≥ nr n−1 ,
W
and it suffices to show the inequality

W
nr 2n−1 + n(n − 1)r 2n−1 + n2 (n − 1)r n−1 ≥ n3 r n ,
which can be rewritten as
n2 r n−1 (r n − nr + n − 1) ≥ 0.
Indeed, by the AM-GM inequality, we get

p
n
r n − nr + n − 1 = r n + 1 + · · · + 1 − nr ≥ n r n · 1 · · · 1 − nr = 0.
The equality holds for a1 = a2 = · · · = an = 1.

P 1.190. If a1 , a2 , . . . , an are positive real numbers, then
b1 b2 bn a1 a2 an
+ + ··· + ≥ + + ··· + ,
a1 a2 an b1 b2 bn
where
1 X
bi = aj, i = 1, 2, · · · , n.
n − 1 j6=i
Solution. Let
a1 + a2 + · · · + an
a= ,
n
L
1 1 1
A= + + ··· + .
.M
a1 a2 an
Using the Cauchy-Schwarz inequality, we have
D
(n − 1)2 1 1 1 1
A
≤ + + ··· + = A− ,
a2 + a3 + · · · + an a2 a3 an a1
n−1 1
PI
M
≤ A− ,
b1 a1
LY
a1 Aa1 − 1
≤ ,
b1 n−1
O
n n
X a i A X n
≤ ai − ,
.M
i=1
bi n − 1 i=1 n−1
n
X a i naA n
≤ − .
W
i=1
bi n−1 n−1
W
Since
n n
X bi 1 X na − ai naA n
= = − ,
W
i=1
a i n − 1 i=1
a i n − 1 n − 1
the conclusion follows. The equality holds for a1 = a2 = · · · = an .
P 1.191. If a1 , a2 , . . . , an are positive real numbers such that
1 1 1
a1 + a2 + · · · + an = + + ··· + ,
a1 a2 an
then
1 1 1
(a) + + ··· + ≥ 1;
1 + (n − 1)a1 1 + (n − 1)a2 1 + (n − 1)an
1 1 1
(b) + + ··· + ≤ 1.
n − 1 + a1 n − 1 + a2 n − 1 + an
Solution. (a) We use the contradiction method. So, assume that
1 1 1
+ + ··· + < 1,
1 + (n − 1)a1 1 + (n − 1)a2 1 + (n − 1)an
and show that
L
1 1 1
a1 + a2 + · · · + an > + + ··· + .
.M
a1 a2 an
Using the substitution
D
1
xi = , i = 1, 2, · · · , n,
A
1 + (n − 1)ai
the hypothesis inequality becomes PI
M
x 1 + x 2 + · · · + x n < 1.
LY
This inequality involves

1 X
1 − x i > (n − 1)bi , bi = i = 1, 2, · · · , n.
O
x j,
n − 1 j6=i
.M
Using this result and the inequality from the preceding P 1.190, we get
n n n
X 1 − xi X bi X x i
a1 + a2 + · · · + an = >
W
≥ .
i=1
(n − 1)x i x
i=1 i
b
i=1 i
W

W
n
X xi 1 1 1
≥ + + ··· + .
b
i=1 i
a1 a2 an
Indeed, we have
n n n
X x i X (n − 1)x i X 1
> = .
b
i=1 i i=1
1 − xi a
i=1 i
The proof is completed. The equality holds for a1 = a2 = · · · = an = 1.
(b) The desired inequality follows from the inequality in (a) by replacing a1 , a2 , . . . , an
with 1/a1 , 1/a2 , . . . , 1/an , respectively.
L
D
.M
A
PI
M
LY
O
.M
W
W
W
Chapter 2
Symmetric Nonrational Inequalities
L
.M
2.1 Applications
D
A
Xp Æ
PI
a2 − a b + b2 ≤ 6(a2 + b2 + c 2 ) − 3(a b + bc + ca).
M
LY

O
v
p p p t a2 + b2 + c 2
a2 − a b + b2 + b2 − bc + c 2 + c 2 − ca + a2 ≤ 3 .
.M
2
W

W
v v v
t 2 t 2 t 2 p
a2 + b2 − a b + b2 + c 2 − bc + c 2 + a2 − ca ≥ 2 a2 + b2 + c 2 .
W
3 3 3

Xp Æ
a2 + a b + b2 ≥ 4(a2 + b2 + c 2 ) + 5(a b + bc + ca).

Xp Æ
a2 + a b + b2 ≤ 5(a2 + b2 + c 2 ) + 4(a b + bc + ca).
257

Xp p p
a2 + a b + b2 ≤ 2 a2 + b2 + c 2 + a b + bc + ca.

p p p p p
a2 + 2bc + b2 + 2ca + c 2 + 2a b ≤ a2 + b2 + c 2 + 2 a b + bc + ca.
L
.M
1 1 1 1 2
p +p +p ≥p +p .
a2 + 2bc b2 + 2ca c 2 + 2a b a2 + b2 + c 2 a b + bc + ca
D
A
p p p p p
PI
2a2 + bc + 2b2 + ca + 2c 2 + a b ≤ 2 a2 + b2 + c 2 + a b + bc + ca.
M
LY
p
2.10. Let a, b, c be nonnegative real numbers such that a + b + c = 3. If k = 3 − 1,
O
then XÆ p
a(a + k b)(a + kc) ≤ 3 3.
.M

W
XÆ
W
a(2a + b)(2a + c) ≥ 9.
W

Æ Æ Æ
b2 + c 2 + a(b + c) + c 2 + a2 + b(c + a) + a2 + b2 + c(a + b) ≥ 6.

p p p
(a) a(3a2 + a bc) + b(3b2 + a bc) + c(3c 2 + a bc) ≥ 6;
p p p p
(b) 3a2 + a bc + 3b2 + a bc + 3c 2 + a bc ≥ 3 3 + a bc.
Symmetric Nonrational Inequalities 259

Æ Æ Æ
a (a + 2b)(a + 2c) + b (b + 2c)(b + 2a) + c (c + 2a)(c + 2b) ≥ 9.

Æ Æ Æ p
a + (b − c)2 + b + (c − a)2 + c + (a − b)2 ≥ 3.
L
.M
v v v
t a(b + c) t b(c + a) t c(a + b)
+ + ≥ 2.
a2 + bc b2 + ca c2 + a b
D
A
1 1
PI 1
p +p +p ≥ 1.
M
3 3 3
a2 + 25a + 1 b2 + 25b + 1 c2 + 25c + 1
LY

O
p p p 3
a2 + bc + b2 + ca + c2 + a b ≤ (a + b + c).
.M
2
W

W
p p p p
a2 + 9bc + b2 + 9ca + c 2 + 9a b ≥ 5 a b + bc + ca.
W

XÆ
(a2 + 4bc)(b2 + 4ca) ≥ 5(a b + ac + bc).

XÆ
(a2 + 9bc)(b2 + 9ca) ≥ 7(a b + ac + bc).

XÆ
(a2 + b2 )(b2 + c 2 ) ≤ (a + b + c)2 .

XÆ
(a2 + a b + b2 )(b2 + bc + c 2 ) ≥ (a + b + c)2 .
L
.M
XÆ
(a2 + 7a b + b2 )(b2 + 7bc + c 2 ) ≥ 7(a b + ac + bc).
D
A
X t
v
7
a + ab + b
2 2

7
b + bc + c ≤
2 2

13
(a + b + c)2 .
PI
9 9 12
M
LY

O
v
X t 1 1 61

a + ab + b
2 2 b + bc + c ≤
2 2 (a + b + c)2 .
.M
3 3 60
W

W
a b c
p +p +p ≥ 1.
4b + bc + 4c
2 2 4c + ca + 4a
2 2 4a + a b + 4b2
2
W

a b c a+b+c
p +p +p ≥p .
b2 + bc + c2 c2 + ca + a2 a2 + ab + b2 a b + bc + ca

a b c a+b+c
p +p +p ≤p .
a2 + 2bc b2 + 2ca c2 + 2a b a b + bc + ca

p p p
a3 + b3 + c 3 + 3a bc ≥ a2 a2 + 3bc + b2 b2 + 3ca + c 2 c 2 + 3a b.
a b c
p +p +p ≤ 1.
4a2 + 5bc 4b2 + 5ca 4c 2 + 5a b
L
2.32. Let a, b, c be nonnegative real numbers. Prove that
.M
p p p
a 4a2 + 5bc + b 4b2 + 5ca + c 4c 2 + 5a b ≥ (a + b + c)2 .
D
A
p p p
PI
a a2 + 3bc + b b2 + 3ca + c c 2 + 3a b ≥ 2(a b + bc + ca).
M
LY

p p p
a a2 + 8bc + b b2 + 8ca + c c 2 + 8a b ≤ (a + b + c)2 .
O
.M
a2 + 2bc b2 + 2ca c 2 + 2a b
W
p
p +p +p ≥3 a b + bc + ca.
b2 + bc + c 2 c 2 + ca + a2 a2 + a b + b2
W
W
2.36. Let a, b, c be nonnegative real numbers, no two of which are zero. If k ≥ 1, then
a k+1 b k+1 c k+1 ak + bk + c k

+ + ≤ .
2a2 + bc 2b2 + ca 2c 2 + a b a+b+c

a2 − bc b2 − ca c2 − a b
(a) p +p +p ≥ 0;
3a2 + 2bc 3b2 + 2ca 3c 2 + 2a b
a2 − bc b2 − ca c2 − a b
(b) p +p +p ≥ 0.
8a2 + (b + c)2 8b2 + (c + a)2 8c 2 + (a + b)2
p
2.38. Let a, b, c be positive real numbers. If 0 ≤ k ≤ 1 + 2 2, then
a2 − bc b2 − ca c2 − a b
p +p +p ≥ 0.
ka2 + b2 + c 2 k b2 + c 2 + a2 kc 2 + a2 + b2

p p p
(a2 − bc) b + c + (b2 − ca) c + a + (c 2 − a b) a + b ≥ 0.
L
.M
p p p
(a2 − bc) a2 + 4bc + (b2 − ca) b2 + 4ca + (c 2 − a b) c 2 + 4a b ≥ 0.
D
A
PI
M
v v v
t a3 t b3 t c3
+ + ≥ 1.
LY
a3 + (b + c)3 b3 + (c + a)3 c 3 + (a + b)3

O

.M
v v v
u
1 1 1 1 1 1
t t
(a + b + c) + + ≥1+ 1+ (a2 + b2 + c 2 ) + +
t
.
W
a b c a2 b2 c 2
W
W

v
1 1 1 1 1 1
t
5+ 2(a2 + b2 + c2) + + − 2 ≥ (a + b + c) + + .
a2 b2 c 2 a b c

Æ
2(1 + a bc) + 2(1 + a2 )(1 + b2 )(1 + c 2 ) ≥ (1 + a)(1 + b)(1 + c).
v v v
t a2 + bc t b2 + ca t c2 + a b 1
+ + ≥2+ p .
b2 + c2 c2 + a2 a2 + b2 2

Æ Æ Æ Æ
a(2a + b + c) + b(2b + c + a) + c(2c + a + b) ≥ 12(a b + bc + ca).
L
.M
Æ Æ Æ
(4a + 5b)(4a + 5c) + b (4b + 5c)(4b + 5a) + c (4c + 5a)(4c + 5b) ≥ 27.
D
a
A
PI
M
Æ Æ Æ
LY
O

.M
p p p Æ
2 + 7a b + 2 + 7bc + 2 + 7ca ≥ 3 3(a b + bc + ca).
W

W
Pp
(a) a(b + c)(a2 + bc) ≥ 6;
W
p p
a(b + c) a2 + 2bc ≥ 6 3;
P
(b)
p
a(b + c) (a + 2b)(a + 2c) ≥ 18.
P
(c)

p p p
a bc + 3 + b ca + 3 + c a b + 3 ≥ 6.

p
(b + c) b2 + c 2 + 7bc ≥ 18;
P
(a)
p p
(b + c) b2 + c 2 + 10bc ≤ 12 3.
P
(b)
2.53. Let a, b, c be nonnegative real numbers such then a + b + c = 2. Prove that

p p p p
a + 4bc + b + 4ca + c + 4a b ≥ 4 a b + bc + ca.
L
.M
p p p p
a2 + b2 + 7a b + b2 + c 2 + 7bc + c 2 + a2 + 7ca ≥ 5 a b + bc + ca.
D
A
2.55. If a, b, c are nonnegative real numbers, then PI
M
p p p Æ
a2 + b2 + 5a b + b2 + c 2 + 5bc + c 2 + a2 + 5ca ≥ 21(a b + bc + ca).
LY

O
v
.M
p p p t2
a a2 + 5 + b b2 + 5 + c c 2 + 5 ≥ (a + b + c)2 .
3
W

W
p p p
a 2 + 3bc + b 2 + 3ca + c 2 + 3a b ≥ (a + b + c)2 .
W

v v v
t 2a + bc t 2b + ca t 2c + a b
(a) a +b +c ≥ 3;
3 3 3
v v v
t a(1 + b + c) t b(1 + c + a) t c(1 + a + b)
(b) a +b +c ≥ 3.
3 3 3

Æ Æ Æ
8(a2 + bc) + 9 + 8(b2 + ca) + 9 + 8(c 2 + a b) + 9 ≥ 15.
9
2.60. Let a, b, c be nonnegative real numbers such that a + b + c = 3. If k ≥ , then
8
p p p p
a2 + bc + k + b2 + ca + k + c 2 + a b + k ≥ 3 2 + k.
L
.M
p p p p
a3 + 2bc + b3 + 2ca + c 3 + 2a b ≥ 3 3.
D
A
PI
M
p p p p
a2 + bc b2 + ca c2 + a b 3 2
+ + ≥ .
LY
b+c c+a a+b 2

O
2.63. If a, b, c are nonnegative real numbers, no two of which are zero,then

.M
p p p
bc + 4a(b + c) ca + 4b(c + a) a b + 4c(a + b) 9
+ + ≥ .
b+c c+a a+b 2
W
W

W
p p p
a a2 + 3bc b b2 + 3ca c c 2 + 3a b
+ + ≥ a + b + c.
b+c c+a a+b

v v v
t 2a(b + c) t 2b(c + a) t 2c(a + b)
+ + ≥ 2.
(2b + c)(b + 2c) (2c + a)(c + 2a) (2a + b)(a + 2b)

v s v v s v
t bc ca t ab t bc ca t ab
+ + ≤ 1 ≤ + + .
3a2 + 6 3b2 + 6 3c 2 + 6 6a2 + 3 6b2 + 3 6c 2 + 3
2.67. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. If k > 1, than
a k (b + c) + b k (c + a) + c k (a + b) ≥ 6.
L
2.68. Let a, b, c be nonnegative real numbers such that a + b + c = 2. If 2 ≤ k ≤ 3, than
.M
a k (b + c) + b k (c + a) + c k (a + b) ≤ 2.
D
A
2.69. Let a, b, c be nonnegative real numbers, no two of which are zero. If m > n ≥ 0,
than PI
bm + c m c m + am am + bm
M
(b + c − 2a) + (c + a − 2b) + (a + b − 2c) ≥ 0.
bn + c n c n + an an + bn
LY

O
p p p
a2 − a + 1 + a2 − a + 1 + a2 − a + 1 ≥ a + b + c.
.M
W

W
p p p
16a2 + 9 + 16b2 + 9 + 16b2 + 9 ≥ 4(a + b + c) + 3.
W

p p p
25a2 + 144 + 25b2 + 144 + 25c 2 + 144 ≤ 5(a + b + c) + 24.
2.73. If a, b are positive real numbers such that a b + bc + ca = 3, then

p p p
(a) a2 + 3 + b2 + 3 +
b2 + 3 ≥ a + b + c + 3;
p p p p
(b) a + b + b + c + c + a ≥ 4(a + b + c) + 6.

Æ Æ Æ
(5a2 + 3)(5b2 + 3) + (5b2 + 3)(5c 2 + 3) + (5c 2 + 3)(5a2 + 3) ≥ 24.

v
p p p t 4(a2 + b2 + c 2 ) + 42
a2 + 1 + b2 + 1 + c 2 + 1 ≥ .
3
L
.M
p p p p
(a) a2 + 3 + b2 + 3 + c 2 + 3 ≥ 2(a2 + b2 + c 2 ) + 30;
p p p
D
p
(b) 3a2 + 1 + 3b2 + 1 + 3c 2 + 1 ≥ 2(a2 + b2 + c 2 ) + 30.
A
PI
M
Æ Æ Æ
(32a2 + 3)(32b2 + 3) + (32b2 + 3)(32c 2 + 3) + (32c 2 + 3)(32a2 + 3) ≤ 105.
LY
O

b+c c + a a + b

.M

a − 3 + − 3 + − 3 ≥ 2.
b c
W
2.79. If a, b, c are real numbers such that a bc 6= 0, then

W
b + c c + a a + b

+ + ≥ 2.
W

a b c
2a 2b 2c
x= , y= , z= .
b+c c+a a+b
Prove that
p p p
(a) x + y +z+ xy+
yz + z x ≥ 6;
p p p
x + y + z ≥ 8 + x yz.
p
(b)
2a 2b 2c
x= , y= , z= .
b+c c+a a+b
Prove that p p p
1 + 24x + 1 + 24 y + 1 + 24z ≥ 15.

v v v
7a 7b 7c
L
t t t
+ + ≤ 3.
a + 3b + 3c b + 3c + 3a c + 3a + 3b
D
.M
2.83. If a, b, c are positive real numbers such that a + b + c = 3, then
A
Æ3
Æ
3
Æ
3
p
3
a2 (b2 + c 2 ) + b2 (c 2 + a2 ) + c 2 (a2 + b2 ) ≤ 3 2.
PI
M
LY
1 1 1 1 2
+ + ≥ +p .
a+b b+c c+a a+b+c a b + bc + ca
O
.M
2.85. If a, b ≥ 1, then
W
1 1 1 1
p + ≥p +p .
3a b + 1 2 3a + 1 3b + 1
W
W
2.86. Let a, b, c be positive real numbers such that a ≥ 1 ≥ b ≥ c and a bc = 1. Prove

that
1 1 1 3
p +p +p ≥ .
3a + 1 3b + 1 3c + 1 2
1
2.87. Let a, b, c be positive real numbers such that a + b + c = 3. If k ≥ p , then
2
(a bc)k (a2 + b2 + c 2 ) ≤ 3.
2.88. Let p and q be nonnegative real numbers such that p2 ≥ 3q, and let
v v
t 2p − 2w t 2p + w
g(p, q) = +2 ,
3 3
2p + 2w
 s s
2p − w

 +2 , p2 ≤ 4q
3 3
h(p, q) = p p ,
 p + p + q, p2 ≥ 4q
p

where w =
p
L
p2 − 3q. If a, b, c are nonnegative real numbers such that
.M
a + b + c = p, a b + bc + ca = q,
D
then
p p p
A
(a) a+b+ b+c+ c + a ≥ g(p, q),
with equality for a =

p + 2w
and b = c =
p−w PI
(or any cyclic permutation);
3 3
M
p p p
(b) a + b + b + c + c + a ≤ h(p, q),
LY
p − 2w p+w
with equality for a = and b = c = (or any cyclic permutation) - when
O
3 3
p2 ≤ 4q, and for a = 0, b+c = p and bc = q (or any cyclic permutation) - when p2 ≥ 4q.
.M
2.89. Let a, b, c, d be nonnegative real numbers such that a2 + b2 + c 2 + d 2 = 1. Prove

W
that p p
p p p p p p
W
1 − a + 1 − b + 1 − c + 1 − d ≥ a + b + c + d.
W
2.90. Let a, b, c, d be positive real numbers. Prove that

p
A + 2 ≥ B + 4,
where
1 1 1 1

A = (a + b + c + d) + + + − 16,
a b c d
1 1 1 1

2 2 2 2
B = (a + b + c + d ) 2 + 2 + 2 + 2 − 16.
a b c d
2.91. Let a1 , a2 , . . . , an be nonnegative real numbers such that a1 + a2 + · · · + an = 1.

Prove that
3a1 + 1 + 3a2 + 1 + · · · + 3an + 1 ≥ n + 1.
p p p
2.92. Let 0 ≤ a < b and a1 , a2 , . . . , an ∈ [a, b]. Prove that

p p 2
a1 + a2 · · · + an − n n a1 a2 · · · an ≤ (n − 1)
p
b− a .
L
1 1 1
.M
+p + ··· + p ≥ 1.
1 + (n2 1 + (n2 − 1)a2 1 + (n2 − 1)an
p
− 1)a1
D
A
n
X 1
PI ≥
1
.
1+ 1 + 4n(n − 1)ai
p
2
M
i=1
LY
2.95. If f is a convex function on a real interval I and a1 , a2 , . . . , an ∈ I, then

O
a + a + ··· + a
1 2 n
f (a1 ) + f (a2 ) + · · · + f (an ) + n(n − 2) f ≥
.M
n
≥ (n − 1)[ f (b1 ) + f (b2 ) + · · · + f (bn )],
W
where
1 X
bi = aj, i = 1, 2, · · · , n.
n − 1 j6=i
W
W

n
X 1 1
≤ .
n−1+ (n − 1)2 + 4nai
p
i=1
2
2.97. If a1 , a2 , . . . , an are positive real numbers such that a1 a2 · · · an = 1, then

v
t a1 + a22 + · · · + an2
u 2
a1 + a2 + · · · + an ≥ n − 1 + .
n
2.98. If a1 , a2 , . . . , an are positive real numbers such that a1 a2 · · · an = 1, then

q Æ
(n − 1)(a12 + a22 + · · · + an2 ) + n − n(n − 1) ≥ a1 + a2 + · · · + an .
2.99. Let a1 , a2 , . . . , an (n ≥ 3) be positive real numbers such that a1 a2 · · · an = 1. If

2n − 1
0<p≤ ,
(n − 1)2
then
1 1 1 n
+p +p ≤p .
L
1 + pa1 1 + pa2 1 + pan 1+p
p
.M
2.100. If a1 , a2 , . . . , an (n ≥ 3) are positive real numbers such that a1 a2 · · · an = 1, then
D
A
n Ç
n
2
X X
(n − 1)2 ai4 + 2n − 1 ≥ PI ai ;
i=1 i=1
M
n
n
2
X q X
ai (n − 1)2 ai2 + 2n − 1 ≥ ai .
LY
i=1 i=1
O
2.101. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an ≥ 1. If k > 1,

.M
then
X a1k
≥ 1.
a1k + a2 + · · · + an
W
W
2.102. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an ≥ 1. If

W
−2
≤ k < 1,
n−2
then
X a1k
≤ 1.
a1k + a2 + · · · + an
2.103. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an ≥ 1. If k > 1,

then X a1
≤ 1.
a1 + a2 + · · · + an
k
2.104. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an ≥ 1. If
2
−1 − ≤ k < 1,
n−2
then X a1
≥ 1.
a1k + a2 + · · · + an
2.105. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. If k ≥ 0,

then
L
X 1
≤ 1.
.M
a1 + a2 + · · · + an
k
D
2.106. Let a1 , a2 , . . . , an be nonnegative real numbers such that a1 + a2 + · · · + an ≥ n.
A
If 1 < k ≤ n + 1, then
a1
+
a2
PI
+ ··· +
an
≤ 1.
a1 + a2 + · · · + ank
M
a1k + a2 + · · · + an a1 + a2k + · · · + an
LY
2.107. Let a1 , a2 , . . . , an be nonnegative real numbers such that a1 + a2 + · · · + an ≤ n.

O
If 0 ≤ k < 1, then
.M
1 1 1
+ + ··· + ≥ 1.
a1k + a2 + · · · + an a1 + a2k + · · · + an a1 + a2 + · · · + ank
W
W
2.108. Let a1 , a2 , . . . , an be positive real numbers. If k > 1, then

W
X ak + ak + · · · + ak n(a1k + a2k + · · · + ank )

2 3 n
≤ .
a2 + a3 + · · · + an a1 + a2 + · · · + an
2.2 Solutions

Xp Æ
a2 − a b + b2 ≤ 6(a2 + b2 + c 2 ) − 3(a b + bc + ca).
Solution. By squaring, the inequality becomes as follows

XÆ
2(a b + bc + ca) + 2 (a2 − a b + b2 )(a2 − ac + c 2 ) ≤ 4(a2 + b2 + c 2 ),
L
2
.M
X p p
a2 − a b + b2 − a2 − ac + c 2 ≥ 0.
D
tion).
A
PI
M
LY
v
p p p t a2 + b2 + c 2
a2 − a b + b2 + b2 − bc + c 2 + c 2 − ca + a2 ≤ 3 .
2
O
.M
Solution (by Nguyen Van Quy). Assume that c = min{a, b, c}. Since
b2 − bc + c 2 ≤ b2
W
and
W
c 2 − ca + a2 ≤ a2 ,
W

v
p t a2 + b2 + c 2
a2 − a b + b2 + a + b ≤ 3 .
2

v
(a + b)2
p t
a2 − ab + b2 +a+b≤ (a2 − ab + b2 ) + (1 + k)
k
v
t (1 + k)[(1 + k)(a2 + b2 ) + (2 − k)a b]
= .
k
Choosing k = 2, we get
v v
p t a2 + b2 t a2 + b2 + c 2
a2 − a b + b2 + a + b ≤ 3 ≤3 = 3.
2 2
The equality holds for a = b and c = 0 (or any cyclic permutation).
L
v v v
2 2 2
.M
t t t p
a2 + b2 − a b + b + c − bc + c 2 + a2 − ca ≥ 2 a2 + b2 + c 2 .
2 2
3 3 3
D
A
First Solution. By squaring, the inequality becomes
2
XÆ PI
(3a2 + 3b2 − 2a b)(3a2 + 3c 2 − 2ac) ≥ 6(a2 + b2 + c 2 ) + 2(a b + bc + ca),
M
X p p 2
6(a2 + b2 + c 2 − a b − bc − ca) ≥ 3a2 + 3b2 − 2a b − 3a2 + 3c 2 − 2ac
LY
,
X X (b − c)2 (3b + 3c − 2a)2
O
3 (b − c)2 ≥ p p 2 ,
3a2 + 3b2 − 2a b + 3a2 + 3c 2 − 2ac
.M
 
X (3b + 3c − 2a)2
(b − c)2 1 − p 2  .
W
p
9a + 9b − 6a b + 9a + 9c − 6ac
2 2 2 2
W
Since p Æ
9a2 + 9b2 − 6a b = (3b − a)2 + 8a2 ≥ |3b − a|,
W
p Æ
9a2 + 9c 2 − 6ac = (3c − a)2 + 8a2 ≥ |3c − a|,
2
|3b + 3c − 2a|
X
2
(b − c) 1 − ≥ 0.
|3b − a| + |3c − a|
This is true since
|3b + 3c − 2a| = |(3b − a) + (3c − a)| ≤ |3b − a| + |3c − a|.
The equality holds for a = b = c, and also for b = c = 0 (or any cyclic permutation).
Second Solution. Assume that a ≥ b ≥ c. Write the inequality as

Æ Æ Æ
(a + b)2 + 2(a − b)2 + (b + c)2 + 2(b − c)2 + (a + c)2 + 2(a − c)2 ≥
Æ
≥ 2 3(a2 + b2 + c 2 ).
By Minkowski’s inequality, it suffices to show that
Æ Æ
[(a + b) + (b + c) + (a + c)]2 + 2[(a − b) + (b − c) + (a − c)]2 ≥ 2 3(a2 + b2 + c 2 ),
L
Æ Æ
(a + b + c)2 + 2(a − c)2 ≥ 3(a2 + b2 + c 2 ).
.M
By squaring, the inequality turns into
D
(a − b)(b − c) ≥ 0.
A
PI
M
LY
Xp Æ
a2 + a b + b2 ≥ 4(a2 + b2 + c 2 ) + 5(a b + bc + ca).
O

.M
First Solution. By squaring, the inequality becomes

XÆ
(a2 + a b + b2 )(a2 + ac + c 2 ) ≥ (a + b + c)2 .
W
Using the Cauchy-Schwarz inequality, we get

W
v
b 2 3b2

W
c 2 3c 2
XÆ Xu
(a + a b + b )(a + ac + c ) =
2 2 2 2 a+ + a+ +
t
2 4 2 4
X c 3bc

b
≥ a+ a+ + = (a + b + c)2 .
2 2 4
Second Solution. Assume that a ≥ b ≥ c. By Minkowski’s inequality, we get
Xp XÆ
2 a2 + a b + b2 = 3(a + b)2 + (a − b)2
Æ
≥ 3[(a + b) + (b + c) + (c + a)]2 + [(a − b) + (b − c) + (a − c)]2
Æ
=2 3(a + b + c)2 + (a − c)2 .
3(a + b + c)2 + (a − c)2 ≥ 4(a2 + b2 + c 2 ) + 5(a b + bc + ca),
(a − b)(b − c) ≥ 0.

• Let a, b, c be nonnegative real numbers such that a b + bc + ca = 4. If |k| ≤ 2, then
L
.M
Xp p
a2 + ka b + b2 ≥ 2 a2 + b2 + c 2 + 3k + 2.
For k = −2/3 and k = 1, we get the inequalities in P 2.3 and P 2.4, respectively. For
D
k = −1 and k = 0, we get the inequalities
A
Xp p
a2 − a b + b2 ≥ 2 a2 + b2 + c 2 − 1,
PI
M
Xp p
a2 + b2 ≥ 2 a2 + b2 + c 2 + 2.
LY
O

.M
Xp Æ
a2 + a b + b2 ≤ 5(a2 + b2 + c 2 ) + 4(a b + bc + ca).
W
(Michael Rozenberg, 2008)

W
First Solution (by Vo Quoc Ba Can). Using the Cauchy-Schwarz inequality, we have
X b2 + bc + c 2

W
X p 2 X
b + bc + c
2 2 ≤ (b + c)
b+c
b + bc + c 2
X 2 X a 2
= 2(a + b + c) =2 1+ (b + bc + c 2 )
b+c b+c
X 2a(b2 + bc + c 2 )
= 4(a2 + b2 + c 2 ) + 2(a b + bc + ca) +
b+c

2 2 2
X bc
= 4(a + b + c ) + 2(a b + bc + ca) + 2a b + c −
b+c
X 1
= 4(a2 + b2 + c 2 ) + 6(a b + bc + ca) − 2a bc .
b+c

X 1
4(a2 + b2 + c 2 ) + 6(a b + bc + ca) − 2a bc ≤ 5(a2 + b2 + c 2 ) + 4(a b + bc + ca),
b+c
which is equivalent to Schur’s inequality
X 1
2(a b + bc + ca) ≤ a2 + b2 + c 2 + 2a bc .
b+c
We can prove this inequality by writing it as follows:
X bc

L
(a + b + c)2 ≤ 2 a a+ ,
b+c
.M
X a
(a + b + c)2 ≤ 2(a b + bc + ca) ,
D
b+c
X a
A
X
(a + b + c)2 ≤ a(b + c) .
b+c
PI
Clearly, the last inequality follows from the Cauchy-Schwarz inequality. The equality
holds for a = b = c.
M
Second Solution. Let us denote
LY
p p p
A = b2 + bc + c 2 , B = c 2 + ca + a2 , C= a2 + a b + b2 .
O
Without loss of generality, assume that a ≥ b ≥ c. For b = c = 0, the inequality is clearly

.M
true. Consider further b > 0. By squaring, the inequality becomes

X X X
2 BC ≤ 3 a2 + 3 a b,
W
W
X X X
a2 − ab ≤ (B − C)2 ,
W
X X (b − c)2
(b − c)2 ≤ 2(a + b + c)2 .
(B + C)2
Since
(B + C)2 ≤ 2(B 2 + C 2 ) = 2(2a2 + b2 + c 2 + ca + a b),
X X (b − c)2
(b − c)2 ≤ (a + b + c)2 ,
2a2 + b2 + c 2 + ca + a b
which is equivalent to X
(b − c)2 Sa ≥ 0,
where
(a + b + c)2 −a2 + a b + 2bc + ca
Sa = − 1 = ,
2a2 + b2 + c 2 + ca + a b 2a2 + b2 + c 2 + ca + a b
−b2 + bc + 2ca + a b
Sb = ≥ 0,
2b2 + c 2 + a2 + a b + bc
−c 2 + ca + 2a b + bc
Sc = ≥ 0.
2c 2 + a2 + b2 + bc + ca
According to
X a2
(b − c)2 Sa ≥ (b − c)2 Sa + (a − c)2 S b ≥ (b − c)2 Sa + 2 (b − c)2 S b
L
b
.M
2 Sa Sb

2 a 2
≥ (b − c) Sa + (b − c) S b = a(b − c) + ,
b a b
D
A
Sa S b
+ ≥ 0,
a b PI
M
−b2 + bc + 2ca + a b a2 − a b − 2bc − ca
≥ .
LY
b(2b2 + c 2 + a2 + a b + bc) a(2a2 + b2 + c 2 + ca + a b)

O
Consider the non-trivial case a2 − a b − 2bc − ca ≥ 0. Since

.M
(2a2 + b2 + c 2 + ca + a b) − (2b2 + c 2 + a2 + a b + bc) = (a − b)(a + b + c) ≥ 0,

W
−b2 + bc + 2ca + a b a2 − a b − 2bc − ca

≥ .
W
b a
Indeed,
W
a(−b2 + bc + 2ca + a b) − b(a2 − a b − 2bc − ca) = 2c(a2 + a b + b2 ) > 0.

Xp p p
a2 + a b + b2 ≤ 2 a2 + b2 + c 2 + a b + bc + ca.

First Solution (by Nguyen Van Quy). Assume that a = max{a, b, c}. Since
p p Æ
a2 + a b + b2 + c 2 + ca + a2 ≤ 2[(a2 + a b + b2 ) + (c 2 + ca + a2 )],

p p p p
2 A + b2 + bc + c 2 ≤ 2 X + Y ,
where
1
A = a2 + (b2 + c 2 + a b + ac), X = a2 + b2 + c 2 , Y = a b + bc + ca.
2
L
Write the desired inequality as follows
.M
p p p p
2( A − X ) ≤ Y − b2 + bc + c 2 ,
D
2(A − X ) Y − (b2 + bc + c 2 )
p p ≤p p ,
A
A+ X Y + b2 + bc + c 2
b(a − b) + c(a − c)
p p
A+ X
≤p p
PI
b(a − b) + c(a − c)
Y + b2 + bc + c 2
.
M
Since b(a − b) + c(a − c) ≥ 0, we need to show that
LY
p p p p
A + X ≥ Y + b2 + bc + c 2 .
O
This inequality is true because X ≥ Y and

p
.M
p
A ≥ b2 + bc + c 2 .
Indeed,
W
2(A − b2 − bc − c 2 ) = 2a2 + (b + c)a − (b + c)2 = (2a − b − c)(a + b + c) ≥ 0.

W
W
Second Solution. In the first solution of P 2.5, we have shown that

X p 2 X 1
b2 + bc + c 2 ≤ 4(a2 + b2 + c 2 ) + 6(a b + bc + ca) − 2a bc .
b+c
X 1 p p 2
4(a2 + b2 +c 2 )+6(a b+ bc +ca)−2a bc ≤ 2 a2 + b2 + c 2 + a b + bc + ca ,
b+c
X 1 Æ
2a bc + 4 (a2 + b2 + c 2 )(a b + bc + ca) ≥ 5(a b + bc + ca).
b+c
Since X 1 9 9
≥P = ,
b+c (b + c) 2(a + b + c)
9a bc Æ
+ 4 (a2 + b2 + c 2 )(a b + bc + ca) ≥ 5(a b + bc + ca),
a+b+c
9a bc Æ
+ 4 q(p2 − 2q) ≥ 5q,
p
where
L
p = a + b + c, q = a b + bc + ca.
.M
For p2 ≥ 4q, this inequality is true because 4 q(p2 − 2q) ≥ 5q. Consider further
p
D
3q ≤ p2 ≤ 4q.
A
By Schur’s inequality of third degree,
9a bc
PI
≥ 4q − p2 .
M
p
LY

O
Æ
(4q − p2 ) + 4 q(p2 − 2q) ≥ 5q,
.M
which is Æ
4 q(p2 − 2q) ≥ p2 + q.
W
Indeed,
16q(p2 − 2q) − (p2 + q)2 = (p2 − 3q)(11q − p2 ) ≥ 0.
W
Third Solution. Let us denote

W
p p p
A = b2 + bc + c 2 , B = c 2 + ca + a2 , C = a2 + a b + b2 ,
p p
X = a2 + b2 + c 2 , Y = a b + bc + ca.
By squaring, the inequality becomes
X X
2 BC ≤ 2 a2 + 4X Y,
X
(B − C)2 ≥ 2(X − Y )2 ,
P 2
2
X (b − c)2 (b − c)2
2(a + b + c) ≥ .
(B + C)2 (X + Y )2
Since
B + C ≤ (c + a) + (a + b) = 2a + b + c,
P 2
2
X (b − c)2 (b − c)2
2(a + b + c) ≥ .
(2a + b + c)2 (X + Y )2

P 2
X (b − c)2 (b − c)2
≥P .
(2a + b + c)2 (b − c)2 (2a + b + c)2
L
.M
Therefore, it is enough to prove that
2(a + b + c)2 1
D
≥ ,
(b − c)2 (2a + b + c)2 (X + Y )2
P
A
which is X PI
2(a + b + c)2 (X + Y )2 ≥ (b − c)2 (2a + b + c)2 .
M
We see that
LY
X X X X
(a + b + c)2 (X + Y )2 ≥ a2 + 2 ab a2 + ab
O
X 2 X X X 2
= a2 + 3 ab a2 + 2 ab
.M
X X X
≥ a4 + 3 a b(a2 + b2 ) + 4 a2 b2
and
W
X X
(b − c)2 (2a + b + c)2 = (b − c)2 [4a2 + 4a(b + c) + (b + c)2 ]
W
X X X
=4 a2 (b − c)2 + 4 a(b − c)(b2 − c 2 ) + (b2 − c 2 )2
W
X X X
≤8 a2 b2 + 4 a(b3 + c 3 ) + 2 a4 .
X X X X X X
a4 + 3 a b(a2 + b2 ) + 4 a2 b2 ≥ 4 a2 b2 + 2 a(b3 + c 3 ) + a4 ,

X
a b(a2 + b2 ) ≥ 0.

p p p p p
a2 + 2bc + b2 + 2ca + c 2 + 2a b ≤ a2 + b2 + c 2 + 2 a b + bc + ca.
(Vasile Cîrtoaje and Nguyen Van Quy, 1989)
Solution (by Nguyen Van Quy). Let

p p
X = a2 + b2 + c 2 , Y= a b + bc + ca.
Consider the nontrivial case when no two of a, b, c are zero (Y 6= 0) and write the
inequality as
L
X p
X − a2 + 2bc ≥ 2(X − Y ),
.M
(b − c)2
P
X (b − c)2
p ≥ .
X +Y
D
X + a2 + 2bc
A
X (b − c)2
p ≥P
P
(b − c)2
p
2
.
PI
X + a2 + 2bc (b − c)2 X + a2 + 2bc
M
LY
(b − c)2
P
1
O
p ≥ ,
(b − c) X + a + 2bc +
P
2 2 X Y
.M
which is equivalent to X p
(b − c)2 Y − a2 + 2bc ≥ 0.
W
From p 2
a2 + 2bc
W
Y− ≥ 0.
we get
W
p Y 2 − (a2 + 2bc) (a − b)(c − a)

Y− a2 + 2bc ≥ = .
2Y 2Y
Thus,
X p X (b − c)2 (a − b)(c − a)
(b − c)2 Y − a2 + 2bc ≥
2Y
(a − b)(b − c)(c − a) X
= (b − c) = 0.
2Y
1 1 1 1 2
p +p +p ≥p +p .
a2 + 2bc b2 + 2ca c 2 + 2a b a2 + b2 + c 2 a b + bc + ca
Solution . Let p p
X= a2 + b2 + c 2 , Y= a b + bc + ca.
Consider the nontrivial case when Y > 0 and write the inequality as
X 1 1 1 1

L
p − ≥2 − ,
a2 + 2bc X Y X
.M
(b − c)2
P
X (b − c)2
p p ≥ .
D
Y (X + Y )

a2 + 2bc X + a2 + 2bc
A
X (b − c)2
P
(b − c)2
2 PI
p p ≥P p p .
M
a2 + 2bc X + a2 + 2bc (b − c)2 a2 + 2bc X + a2 + 2bc
LY

O
(b − c)2
P
1
p p ≥ ,
(b − c) a + 2bc X + a + 2bc (X + Y)
P
2 2 2 Y
.M
W
X p
(b − c)2 [X Y − X a2 + 2bc + (a − b)(c − a)] ≥ 0.
W
Since X X
(b − c)2 (a − b)(c − a) = (a − b)(b − c)(c − a) (b − c) = 0,
W

X p
(b − c)2 X Y − a2 + 2bc ≥ 0,
X p
(b − c)2 Y − a2 + 2bc ≥ 0.
We have proved this inequality at the preceding problem P 2.7. The equality holds for
a = b, or b = c, or c = a.

p p p p p
2a2 + bc + 2b2 + ca + 2c 2 + a b ≤ 2 a2 + b2 + c 2 + a b + bc + ca.
Solution. We will apply Lemma below for
X = 2a2 + bc, Y = 2b2 + ca, Z = 2c 2 + a b
and
A = B = a2 + b2 + c 2 , C = a2 + b2 + c 2 .
L
We have
X + Y + Z = A+ B + C
.M
and
A = B ≥ C.
D
A
Without loss of generality, assume that
a ≥ b ≥ c. PI
M
By Lemma, it suffices to show that
LY
max{X , Y, Z} ≥ A, min{X , Y, Z} ≤ C.
O
Indeed, we have
max{X , Y, Z} − A ≥ X − A = (a2 − b2 ) + c(b − c) ≥ 0,

.M
min{X , Y, Z} − C ≤ Z − C = c(2c − a − b) ≤ 0.
W
Equality holds for a = b = c.

W
Lemma. If X , Y, Z and A, B, C are positive real numbers such that
X + Y + Z = A + B + C,
W
max{X , Y, Z} ≥ max{A, B, C}, min{X , Y, Z} ≤ min{A, B, C},

then p p p p p p
X+ Y+ Z≤ A + B + C.
Proof. On the assumption that X ≥ Y ≥ Z and A ≥ B ≥ C, we have
X ≥ A, Z ≤ C,
and hence
p p p p p p p p p p p p
X + Y + Z − A− B − C = ( X − A ) + ( Y − B ) + ( Z − C )
X −A Y −B Z −C X −A Y −B Z −C
≤ p + p + p ≤ p + p + p
2 A 2 B 2 C 2 B 2 B 2 C
C −Z Z −C 1 1

= p + p = (C − Z) p − p ≤ 0.
2 B 2 C 2 B 2 C
Remark. This Lemma is a particular case of Karamata’s inequality.
p
P 2.10. Let a, b, c be nonnegative real numbers such that a + b + c = 3. If k = 3 − 1,
L
then p
.M
XÆ
a(a + k b)(a + kc) ≤ 3 3.
D
A
XÆ
a(a + k b)(a + kc) ≤
rX X
a PI
(a + k b)(a + kc) .
M
LY
rX
(a + k b)(a + kc) ≤ a + b + c,
O
which is an identity. The equality holds for a = b = c = 1, and also for a = 3 and
.M

W

W
XÆ
W
a(2a + b)(2a + c) ≥ 9.

X Æ Æ
a(2a + b)(2a + c) − a 3(a + b + c) ≥ 0,
X
(a − b)(a − c)Ea ≥ 0,
where p
a
Ea = p p .
(2a + b)(2a + c) + 3a(a + b + c)
Assume that a ≥ b ≥ c. Since (c − a)(c − b)Ec ≥ 0, it suffices to show that
(a − c)Ea ≥ (b − c)E b ,
Æ Æ Æ
(a − b) 3a b(a + b + c) + (a − c) a(2b + c)(2b + a) ≥ (b − c) b(2a + b)(2a + c).
This is true if
Æ Æ
(a − c) a(2b + c)(2b + a) ≥ (b − c) b(2a + b)(2a + c).
L
.M
For the non-trivial case b > c, we have
p
D
a−c a a
≥ ≥p .
b−c b b
A
Therefore, it is enough to show that PI
M
a2 (2b + c)(2b + a) ≥ b2 (2a + b)(2a + c).
LY

O
a2 (2a b + 2bc + ca) ≥ b2 (2a b + bc + 2ca).

.M
It is true if
a(2a b + 2bc + ca) ≥ b(2a b + bc + 2ca).
W
Indeed,
W
a(2a b + 2bc + ca) − b(2a b + bc + 2ca) = (a − b)(2a b + bc + ca) ≥ 0.

W
The equality holds for a = b = c = 1, and also for a = 0 and b = c = 3/2 (or any cyclic
permutation).

Æ Æ Æ
b2 + c 2 + a(b + c) + c 2 + a2 + b(c + a) + a2 + b2 + c(a + b) ≥ 6.
Solution. Denote
A = b2 + c 2 + a(b + c), B = c 2 + a2 + b(c + a), C = a2 + b2 + c(a + b).
First Solution. Write the inequality in the homogeneous form

p p p
A + B + C ≥ 2(a + b + c).

Xp X X
2 BC ≥ 2 a2 + 6 bc,
L
X X p p 2
.M
(a − b)2 ≥ B− C ,
X
(b − c)2 Sa ≥ 0,
D
where
A
(b + c − a)2
Sa = 1 − p p PI.
( B + C)2
M
Since
(b + c − a)2 a(a + 3b + 3c)
Sa ≥ 1 − = ≥ 0, S b ≥ 0, Sc ≥ 0,
LY
B+C B+C
the conclusion follows. The equality holds for a = b = c = 1, and also for a = 3 and
O

.M
Second Solution. Write the original inequality as follows

Xp
( A − b − c) ≥ 0,
W
X c(a − b) + b(a − c)
W
p ≥ 0,
A+ b + c
W
X c(a − b) X c(b − a)
p + p ≥ 0,
A+ b + c B+c+a
X c(a − b)[a − b − (pA − pB)]
p p ≥ 0.
( A + b + c)( B + c + a)
p p
(a − b)[a − b + ( B − A)] ≥ 0.
Indeed,
p p a+b−c

2
(a − b)[a − b + ( B − A)] = (a − b) 1 + p p ≥ 0,
B+ A
because, for the nontrivial case a + b − c < 0, we have
a+b−c a+b−c
1+ p p >1+ > 0.
B+ A c+c
16
Open Generalization. Let a, b, c be nonnegative real numbers. If 0 < k ≤ , then
9
XÆ
(b + c)2 + k(a b − 2bc + ca) ≥ 2(a + b + c).
16
Notice that if k = , then the equality holds for a = b = c = 1, for a = 0 and b = c (or
9
L
any cyclic permutation), and for b = c = 0 (or any cyclic permutation).
D
.M
A
(a)
p
a(3a2 + a bc) +
p
PI
b(3b2 + a bc) +
p
c(3c 2 + a bc) ≥ 6;
p p p p
M
(b) 3a2 + a bc + 3b2 + a bc + 3c 2 + a bc ≥ 3 3 + a bc.
LY
(Lorian Saceanu, 2015)
Solution. (a) Write the inequality in the homogeneous form

O
X Æ
3 a (a + b)(a + c) ≥ 2(a + b + c)2 .
.M

W
X X 3 X p p 2
a2 − ab ≥ a a+b− a+c ,
2
W
X X a(b − c)2
(b − c)2 ≥ 3
W
p p 2 ,
a+b+ a+c
X
(b − c)2 Sa ≥ 0,
where
3a
Sa = 1 − p p 2 .
a+b+ a+c
Since
3a
Sa ≥ 1 − p p 2 > 0, S b > 0, Sc > 0,
a+ a
the inequality is true. The equality holds for a = b = c = 1.
Second Solution. By Hölder’s inequality, we have
( a)3
P
X Æ 2 27
a (a + b)(a + c) ≥ P a =P a .
(a + b)(a + c) (a + b)(a + c)

X a 3
≤ .
(a + b)(a + c) 4
L
This inequality has the homogeneous form
.M
X a 9
≤ ,
(a + b)(a + c) 4(a + b + c)
D
A
X
PI
a(b − c)2 ≥ 0.
M
(b) By squaring, the inequality becomes
LY
X XÆ
3 a2 + 2 (3b2 + a bc)(3c 2 + a bc) ≥ 27 + 6a bc.
O

.M
Æ
(3b2 + a bc)(3c 2 + a bc) ≥ 3bc + a bc.
W

W
X X
3 a2 + 6 bc + 6a bc ≥ 27 + 6a bc,
W
which is an identity. The equality holds for a = b = c = 1, and also for a = 0, or b = 0,

or c = 0.

Æ Æ Æ
First Solution. Write the inequality as follows:

X Æ
a (a + 2b)(a + 2c) ≥ 3(a b + bc + ca),
X X 1 X p p 2
a2 − ab ≥ a a + 2b − a + 2c ,
2
X X a(b − c)2
(b − c)2 ≥ 4 p p 2 ,
a + 2b + a + 2c
X
(b − c)2 Sa ≥ 0,
where
L
4a
Sa = 1 − p 2 .
.M
p
a + 2b + a + 2c
Since
D
4a
Sa > 1 − p p 2 = 0, S b > 0, Sc > 0,
A
a+ a
PI
the inequality is true. The equality holds for a = b = c = 1.
Second Solution. We use the AM-GM inequality to get
M
X Æ X 2a(a + 2b)(a + 2c) X 2a(a + 2b)(a + 2c)
LY
a (a + 2b)(a + 2c) = ≥
(a + 2b) + (a + 2c)
p
2 (a + 2b)(a + 2c)
1
O
X
= a(a + 2b)(a + 2c).
a+b+c
.M

X
a(a + 2b)(a + 2c) ≥ 9(a + b + c).
W
Write this inequality in the homogeneous form

W
X
a(a + 2b)(a + 2c) ≥ 3(a + b + c)(a b + bc + ca),
W
which is equivalent to Schur’s inequality of degree three
a3 + b3 + c 3 + 3a bc ≥ a b(a + b) + bc(b + c) + ca(c + a).

Æ Æ Æ p
a + (b − c)2 + b + (c − a)2 + c + (a − b)2 ≥ 3.
(Phan Thanh Nam, 2007)

Solution. By squaring, the inequality becomes

XÆ
[a + (b − c)2 ][b + (a − c)2 ] ≥ 3(a b + bc + ca).
Applying the Cauchy-Schwarz inequality, it suffices to show that

Xp X
ab + (b − c)(a − c) ≥ 3(a b + bc + ca).
This is equivalent to the homogeneous inequality

X X p X
a ab + a2 ≥ 4(a b + bc + ca).
L
p p p
Making the substitution x = a, y = b, z = c, the inequality turns into
.M
X X X X
x2 xy + x4 ≥ 4 x 2 y 2,
D
A
X
x4 +
X
x y(x 2 + y 2 ) + x yz
X
x ≥4 PI
X
x 2 y 2.
M
Since X X
4 x2 y2 ≤ 2 x y(x 2 + y 2 ),
LY

O
X X X
x 4 + x yz x≥ x y(x 2 + y 2 ),
.M
1
which is just Schur’s inequality of degree four. The equality holds for a = b = c = ,
3
W
1
and for a = 0 and b = c = (or any cyclic permutation).
2
W
W
v v v
t a(b + c) t b(c + a) t c(a + b)
+ + ≥ 2.
a2 + bc b2 + ca c2 + a b
Solution. Using the AM-GM inequality gives

v
t a(b + c) a(b + c) 2a(b + c) 2a(b + c)
=p ≥ 2 = .
a2 + bc (a + bc)(ab + ac)
2 (a + bc) + (a b + ac) (a + b)(a + c)
a(b + c) b(c + a) c(a + b)

+ + ≥ 1,
(a + b)(a + c) (b + c)(b + a) (c + a)(c + b)
a(b + c)2 + b(c + a)2 + c(a + b)2 ≥ (a + b)(b + c)(c + a),
4a bc ≥ 0.
L
.M
D
1 1 1
A
p
3
+p
3
+p
3
≥ 1.
a2 + 25a + 1 b2 + 25b + 1 PI c2 + 25c + 1
M
Solution. Replacing a, b, c by a3 , b3 , c 3 , respectively, we need to show that a bc = 1
yields
LY
1 1 1
p
3
+p 3
+p 3
≥ 1.
a6 + 25a3 + 1 b6 + 25b3 + 1 c 6 + 25c 3 + 1
O
We first show that

1 1
.M
p ≥ 2 .
3
a + 25a + 1
6 3 a +a+1
W
(a2 + a + 1)3 ≥ a6 + 25a3 + 1,

W
which is true since
(a2 + a + 1)3 − (a6 + 25a3 + 1) = 3a(a − 1)2 (a2 + 4a + 1) ≥ 0.

W
1 1 1
+ + ≥ 1.
a2 + a + 1 b2 + b + 1 b2 + b + 1
Putting
yz zx xy
a= , b= , c= , x, y, z > 0
x2 y2 z2
X x4
≥ 1.
x 4 + x 2 yz + y 2 z 2
Indeed, the Cauchy-Schwarz inequality gives

P 2 2
x + 2 x2 y2
P 4 P
X x4 x
≥P =P
x 4 + x 2 yz + y 2 z 2 (x 4 + x 2 yz + y 2 z 2 ) x 4 + x yz x + x 2 y 2
P P
P 2 2 P
x y − x yz x
=1+ P ≥ 1.
x 4 + x yz x + x 2 y 2
P P
L
.M
p p p 3
a2 + bc + b2 + ca + c2 + a b ≤ (a + b + c).
D
2
A
PI
Solution. Without loss of generality, assume that a ≥ b ≥ c. Since the equality occurs
for a = b and c = 0, we use the inequalities
M
p c
a2 + bc ≤ a +
LY
2
and
O
p p Æ
b2 + ca + c2 + a b ≤ 2(b2 + ca) + 2(c 2 + a b).
.M

Æ a + 3b + 2c
2(b2 + ca) + 2(c 2 + a b) ≤ .
W
2
By squaring, this inequality becomes
W
a2 + b2 − 4c 2 − 2a b + 12bc − 4ca ≥ 0,
W
(a − b − 2c)2 + 8c(b − c) ≥ 0.

p p p p
a2 + 9bc + b2 + 9ca + c 2 + 9a b ≥ 5 a b + bc + ca.

First Solution (by Nguyen Van Quy). Assume that c = min{a, b, c}. Since the equality
occurs for a = b and c = 0, we use the inequality
p p
c 2 + 9a b ≥ 3 a b.
Also, by Minkowski’s inequality, we have

s
p p p p 2
a2 + 9bc + b2 + 9ca ≥ (a + b)2 + 9c a+ b .

s
p
L
p 2 p p
(a + b)2 + 9c a + b ≥ 5 a b + bc + ca − 3 a b.
.M
D
p Æ
(a + b)2 + 18c a b + 30 a b(a b + bc + ca) ≥ 34a b + 16c(a + b).
A
Since
c(a + b)
2
PI
c(a + b)(3a b − ac − bc)
a b(a b + bc + ca) − a b + =
M
≥ 0,
3 9
p
LY
it suffices to show that f (c) ≥ 0 for 0 ≤ c ≤ a b, where

O
p
f (c) = (a + b)2 + 18c a b + [30a b + 10c(a + b)] − 34a b − 16c(a + b)
p
= (a + b)2 − 4a b + 6c(3 a b − a − b).
.M
p
Since f (c) is a linear function, we only need to prove that f (0) ≥ 0 and f ( a b) ≥ 0.
W
We have
f (0) = (a − b)2 ≥ 0,
W
p p p
f ( a b) = (a + b)2 + 14a b − 6(a + b) a b ≥ (a + b)2 + 9a b − 6(a + b) a b
W
p 2
= a + b − 3 a b ≥ 0.
Second Solution. Assume that c = min{a, b, c}. By squaring, the inequality becomes
X XÆ X
a2 + 2 (a2 + 9bc)(b2 + 9ca) ≥ 16 a b,
X Æ p p p X
a2 + 2 (a2 + 9bc)(b2 + 9ca) + 2 c 2 + 9a b a2 + 9bc + b2 + 9ca ≥ 16 a b.
The Cauchy-Schwarz inequality gives
Æ p
(a2 + 9bc)(b2 + 9ca) ≥ a b + 9c a b.
In addition, Minkowski’s inequality gives

s
p p p p 2
a2 + 9bc + b2 + 9ca ≥ (a + b)2 + 9c a + b ≥ a + b + 4c,
hence p p p p
c 2 + 9a b a2 + 9bc + b2 + 9ca ≥ 3 a b (a + b + 4c).
Therefore, it suffices to show that f (c) ≥ 0, where
p p X
f (c) = a2 + b2 + 2(a b + 9c a b) + 6 a b (a + b + 4c) − 16 ab
p p
= a2 + b2 − 14a b + 6(a + b) a b + c[42 a b − 16(a + b)].
L
p
.M
Since fp(c) is a linear function and 0 ≤ c ≤ a b, it is sufficient to show that f (0) ≥ 0
and f ( a b) ≥ 0. We have
D
p p p 2
f (0) = (a − b)2 + 6 ab a− b ≥0
A
and
p p
PI p
f ( a b) = a2 + b2 + 28a b − 10(a + b) a b ≥ (a + b)2 + 25a b − 10(a + b) a b
M
p 2
LY

= a + b − 5 a b ≥ 0.
O
.M

XÆ
(a2 + 4bc)(b2 + 4ca) ≥ 5(a b + ac + bc).
W

W
First Solution (by Michael Rozenberg). Assume that a ≥ b ≥ c. For b = c = 0, the

W
inequality is trivial. Assume further that b > 0 and write the inequality as follows:
X Æ
[ (b2 + 4ca)(c 2 + 4a b) − bc − 2a(b + c)] ≥ 0,
X (b2 + 4ca)(c 2 + 4a b) − (bc + 2a b + 2ac)2

p ≥ 0,
(b2 + 4ca)(c 2 + 4a b) + bc + 2a(b + c)
X
(b − c)2 Sa ≥ 0,
where
a(b + c − a) Æ
Sa = , A = (b2 + 4ca)(c 2 + 4a b) + bc + 2a(b + c),
A
b(c + a − b) Æ
Sb = , B= (c 2 + 4a b)(a2 + 4bc) + ca + 2b(c + a),
B
c(a + b − c) Æ
Sc = , C = (a2 + 4bc)(b2 + 4ac) + a b + 2c(a + b).
C
Since S b ≥ 0 and Sc ≥ 0, we have
X a2
b
bSa aS b

a
= (b − c)2 + .
b a b
L
.M
bSa aS b
+ ≥ 0,
a b
D
b(b + c − a) a(c + a − b)
A
+ ≥ 0.
A B
Since
PI
M
b(b + c − a) a(c + a − b) b(b − a) a(a − b) (a − b)(aA − bB)
+ ≥ + = ,
LY
A B A B AB
O
aA − bB ≥ 0.
.M
Indeed,
p p p
aA − bB = c 2 + 4a b a b2 + 4ca − b a2 + 4bc + 2(a − b)(a b + bc + ca)
W
p
4c(a3 − b3 ) c 2 + 4a b
W
= p p + 2(a − b)(a b + bc + ca) ≥ 0.

a b2 + 4ca + b a2 + 4bc
W
The equality holds for a = b = c, and also for a = b and c = 0 (or any cyclic permuta-
tion).
Second Solution (by Nguyen Van Quy). Write the inequality as
p p p 2
a2 + 4bc + b2 + 4ca + c 2 + 4a b ≥ a2 + b2 + c 2 + 14(a b + bc + ca),
p p p Æ
a2 + 4bc + b2 + 4ca + c 2 + 4a b ≥ a2 + b2 + c 2 + 14(a b + bc + ca).
Assume that c = min{a, b, c}. For t = 2c, the inequality (b) in Lemma below becomes
p p Æ
a2 + 4bc + b2 + 4ca ≥ (a + b)2 + 8(a + b)c.

Æ p Æ
(a + b)2 + 8(a + b)c + c 2 + 4a b ≥ a2 + b2 + c 2 + 14(a b + bc + ca).

Æ
[(a + b)2 + 8(a + b)c] (c 2 + 4a b) ≥ 4a b + 3(a + b)c,
2(a + b)c 3 − 2(a + b)2 c 2 + 2a b(a + b)c + a b(a + b)2 − 4a2 b2 ≥ 0,

2(a + b)(a − c)(b − c)c + a b(a − b)2 ≥ 0.
L
Lemma. Let a,b and t be nonnegative numbers such that
.M
t ≤ 2(a + b).
D
Then,
A
p
(a) (a2 + 2bt)(b2 + 2at) ≥ a b + (a + b)t;
(b)
p
a2 + 2bt +
p
b2 + 2at ≥
p PI
(a + b)2 + 4(a + b)t.
M
Proof. (a) By squaring, the inequality becomes
LY
(a − b)2 t[2(a + b) − t] ≥ 0,
O
which is clearly true.

.M
(b) By squaring, this inequality turns into the inequality in (a).

W
W

W
XÆ
(a2 + 9bc)(b2 + 9ca) ≥ 7(a b + ac + bc).
Solution (by Nguyen Van Quy). We see that the equality holds for a = b and c = 0.
Without loss of generality, assume that c = min{a, b, c}. For t = 4c, the inequality (a)
in Lemma from the preceding P 2.20 becomes
Æ
(a2 + 8bc)(b2 + 8ca) ≥ a b + 4(a + b)c.
Thus, we have Æ
(a2 + 9bc)(b2 + 9ca) ≥ a b + 4(a + b)c,
and also,
p p p p Æ
4
c 2 + 9a b a2 + 9bc + b2 + 9ca ≥ 3 a b · 2 (a2 + 9bc)(b2 + 9ca)
p Æ Æ
≥ 6 a b · a b + 4(a + b)c = 3 4a2 b2 + 16a bc(a + b)
Æ
≥ 3 4a2 b2 + 4a bc(a + b) + c 2 (a + b)2 = 3(2a b + bc + ca).
Therefore,
XÆ
(a2 + 9bc)(b2 + 9ca) ≥ (a b + 4bc + 4ca) + 3(2a b + bc + ca)
= 7(a b + bc + ca).
L
.M
D
A
Æ Æ Æ PI
(a2 + b2 )(b2 + c 2 ) + (b2 + c 2 )(c 2 + a2 ) + (c 2 + a2 )(a2 + b2 ) ≤ (a + b + c)2 .
M
LY
Solution. Without loss of generality, assume that a = min{a, b, c}. Let us denote
O
a a
y= + b, z= + c.
2 2
.M
Since
a2 + b2 ≤ y 2 , b2 + c 2 ≤ y 2 + z 2 , c 2 + a2 ≤ z 2 ,
W

W
Æ
yz + ( y + z) y 2 + z 2 ≤ ( y + z)2 .
W
This is true since

Æ y 2z2
y 2 + yz + z 2 − ( y + z) y 2 + z 2 = ≥ 0.
y 2 + yz + z 2 + ( y + z) y 2 + z 2
p
The equality holds for a = b = 0 (or any cyclic permutation).

XÆ
(a2 + a b + b2 )(b2 + bc + c 2 ) ≥ (a + b + c)2 .
b 2 3b2 c 2 3c 2

2 2 2 2
(a + a b + b )(a + ac + c ) = a + + a+ +
2 4 2 4
c 3bc a(b + c)

b
≥ a+ a+ + = a2 + + bc.
2 2 4 2
Then,
X a(b + c)
XÆ
2
(a + a b + b )(a + ac + c ) ≥
2 2 2 2 a + + bc = (a + b + c)2 .
2
L
D
.M
A
XÆ
(a2 + 7a b + b2 )(b2 + 7bc + c 2 ) ≥ 7(a b + ac + bc).
M
First Solution. Without loss of generality, assume that c = min{a, b, c}. We see that the
LY
equality holds for a = b and c = 0. Since

Æ
O
(a2 + 7ac + c 2 )(b2 + 7bc + c 2 ) ≥ (a + 2c)(b + 2c) ≥ a b + 2c(a + b),

.M

p p p
a2 + 7a b + b2 a2 + 7ac + b2 + 7bc ≥ 6a b + 5c(a + b).
W
By Minkowski’s inequality, we have

W
s
p p p p 2
a2 + 7ac + b2 + 7bc ≥ (a + b)2 + 7c a+ b
W
v
t 28a bc
≥ (a + b)2 + 7c(a + b) + .
a+b
28a bc

2 2 2
(a + 7a b + b ) (a + b) + 7c(a + b) + ≥ (6a b + 5bc + 5ca)2 .
a+b
1
Due to homogeneity, we may assume that a + b = 1. Let us denote d = a b, d ≤ .
4
Since
2a b
c≤ = 2d,
a+b
1
we need to show that f (c) ≥ 0 for 0 ≤ c ≤ 2d ≤ , where
2
f (c) = (1 + 5d)(1 + 7c + 28cd) − (6d + 5c)2 .
Since f (c) is concave, it suffices to show that f (0) ≥ 0 and f (2d) ≥ 0. Indeed,
f (0) = 1 + 5d − 36d 2 = (1 − 4d)(1 + 9d) ≥ 0
and
f (2d) = (1 + 5d)(1 + 14d + 56d 2 ) − 256d 2 ≥ (1 + 4d)(1 + 14d + 56d 2 ) − 256d 2
L
.M
= (1 − 4d)(1 + 22d − 56d 2 ) ≥ d(1 − 4d)(22 − 56d) ≥ 0.
D
Second Solution. We will use the inequality
A
Æ
x 2 + 7x y + y 2 ≥ x + y +
2x y
x+y
PI
, x, y ≥ 0,
M
which, by squaring, reduces to
LY
x y(x − y)2 ≥ 0.
We have
O
X 2a b 2ac
XÆ
.M
(a2 + 7a b + b2 )(a2 + 7ac + c2) ≥ a+b+ a+c+

a+b a+c
X 2a2 b X 2a2 c X 2a bc
W
X X
≥ a2 + 3 ab + + + .
a+b a+c a+b
W
Since
X 2a2 b X 2a2 c X 2a2 b X 2b2 a X
+ = + =2 ab
W
a+b a+c a+b b+a

and
X 2a bc 18a bc 9a bc
≥P = ,
a+b (a + b) a+b+c
X 9a bc X
a2 + ≥2 a b,
a+b+c
which is just Schur’s inequality of degree three.

v
X t 7 7 13

a2 + a b + b2 b2 + bc + c 2 ≤ (a + b + c)2 .
9 9 12
Solution (by Nguyen Van Quy). Without loss of generality, assume that c = min{a, b, c}.
It is easy to see that the equality holds for a = b = 1 and c = 0. By the AM-GM inequality,
the following inequality holds for any k > 0:
v v v
t 7 t 7 t 7
12 a + a b + b
2 2 a + ac + c + b + bc + c ≤
2 2 2 2
L
9 9 9
.M
v v 2
36 2 7 7 7
t t
≤ a + a b + b2 + k a2 + ac + c 2 + b2 + bc + c 2 .
k 9 9 9
D
We can use this inequality to prove the original inequality only if
A
2
36 2 7
k

a + ab + b = k
9
2
v
t 7
v
tPI 7
a2 + ac + c 2 + b2 + bc + c 2
9 9
M
for a = b = 1 and c = 0. This necessary condition if satisfied for k = 5. Therefore, it
LY

v
7 7 36 2 7

O
t
2
12 a + ac + c
2 2 b + bc + c +
2 2 a + ab + b
9 9 5 9
.M
v v 2
t 7 t 7
+5 a2 + ac + c 2 + b2 + bc + c 2 ≤ 13(a + b + c)2 .
9 9
W
W
v
7 7 4(a + b)2 + 94a b 199c(a + b)
t
22 a + ac + c
2 2 b + bc + c ≤
2 2 + 3c 2 + .
W
9 9 5 9
Since
v v
7 7 16 16
t t
2 a2 + ac + c 2 b + bc + c ≤ 2
2 2 a +
2 ac b +
2 bc
9 9 9 9
v
16 16
t
=2 a b+ c ·b a+ c
9 9
16 16

≤a b+ c +b a+ c
9 9
16c(a + b)
= 2a b + ,
9
we only need to prove that
8c(a + b) 4(a2 + b2 ) + 102a b 199c(a + b)

22 a b + ≤ + 3c 2 + .
9 5 9
4(a − b)2 23c(a + b)

+ + 3c 2 ≥ 0.
5 9
Thus, the proof is completed. The equality holds for a = b and c = 0 (or any cyclic
permutation).
L
.M
D
v
A
X t 1 1 61

a + ab + b
2 2 b + bc + c ≤
2 2 (a + b + c)2 .
3 3 60
M
Solution (by Nguyen Van Quy). Without loss of generality, assume that c = min{a, b, c}.
LY
It is easy to see that the equality holds for c = 0 and 11(a2 + b2 ) = 38a b. By the AM-GM
inequality, the following inequality holds for any k > 0:
O
v v v
t 1 t 1 t 1
60 a2 + a b + b2 a2 + ac + c 2 + b2 + bc + c 2 ≤
.M
3 3 3
v v 2
36 2 1 1 1

W
t t
2
≤ a + a b + b + 25k a + ac + c + b + bc + c
2 2 2 2 .
k 3 3 3
W
We can use this inequality to prove the original inequality only if the equality
W
v v 2
36 2 1 1 1
t t
2
a + a b + b = 25k a2 + ac + c 2 + b2 + bc + c 2
k 3 3 3
holds for c = 0 and 11(a2 + b2 ) = 38a b. This necessary condition if satisfied for k = 1.
v
1 1 1
t
2 2
60 a + ab + b
2 2 b + bc + c + 36 a + a b + b
2 2
3 3 3
v v 2
t 1 t 1
+25 a2 + ac + c 2 + b2 + bc + c 2 ≤ 61(a + b + c)2 ,
3 3
v
1 1 31c(a + b)
t
10 a + ac + c
2 2 b + bc + c ≤ 10a b + c 2 +
2 2 .
3 3 3
Since
v v
1 1 4 4
t t
2 a2 + ac + c 2 b2 + bc + c 2 ≤ 2 a2 + ac b2 + bc
3 3 3 3
v
4 4
t
=2 a b+ c ·b a+ c
3 3
4 4

L
≤a b+ c +b a+ c
3 3
.M
4c(a + b)
= 2a b + ,
3
D
we only need to prove that
A
2c(a + b) 31c(a + b)

10 a b + ≤ 10a b + c 2 + PI .
3 3
M
3c 2 + 11c(a + b) ≥ 0.
LY
Thus, the proof is completed. The equality holds for 11(a2 + b2 ) = 38a b and c = 0 (or
O

.M

W
a b c
p +p +p ≥ 1.
W
4b + bc + 4c
2 2 4c + ca + 4a
2 2 4a + a b + 4b2
2

W
Solution. By Hölder’s inequality, we have

P 3
a + 3 a b(a + b) + 6a bc
2 P 3 P
a
X
a
p ≥P = .
a(4b2 + bc + 4c 2 ) 4 a b(a + b) + 3a bc
P
4b2 + bc + 4c 2
X X
a3 + 3a bc ≥ a b(a + b),
which is Schur’s inequality of degree three. The equality holds for a = b = c, and also
for a = 0 and b = c (or any cyclic permutation).
a b c a+b+c
p +p +p ≥p .
b2 + bc + c2 c2 + ca + a2 a2 + ab + b2 a b + bc + ca

2 P 3 P 2
a a
X
a
p ≥P = P ,
b2 + bc + c 2 a(b2 + bc + c 2 ) ab
from which the desired inequality follows. The equality holds for a = b = c, and also
L
for a = 0 and b = c (or any cyclic permutation).
D
.M
A
p
a2
a
+ 2bc
+p
b2
b
+ 2ca
+p
c2
c
+ 2a b
PI ≤p
a+b+c
a b + bc + ca
.
M
(Ho Phu Thai, 2007)
LY
Solution. Without loss of generality, assume that a ≥ b ≥ c.

O

c c
p ≤p ,
.M
c2 + 2a b a b + bc + ca
a+b
W
a b
p +p ≤p ,
a2 + 2bc b2 + 2ca a b + bc + ca
W
W
p p p p
a( a2 + 2bc − a b + bc + ca) b( a b + bc + ca − b2 + 2ca)
p ≥ p .
a2 + 2bc b2 + 2ca
Since p p
a2 + 2bc − a b + bc + ca ≥ 0
and
a b
p ≥p ,
a2 + 2bc b2 + 2ca
p p p p
a2 + 2bc − a b + bc + ca ≥ a b + bc + ca − b2 + 2ca,
p p p
a2 + 2bc + b2 + 2ca ≥ 2 a b + bc + ca.
Using the AM-GM inequality, it suffices to show that
(a2 + 2bc)(b2 + 2ca) ≥ (a b + bc + ca)2 ,
c(a − b)2 (2a + 2b − c) ≥ 0.
L
.M
tion).
D
X 2
a X X a
≤
A
p a .
a2 + 2bc a2 + 2bc
PI
M
X a a+b+c
≤ .
a2 + 2bc a b + bc + ca
LY
O
X 1 1

a − 2 ≥ 0,
a b + bc + ca a + 2bc
.M
X a(a − b)(a − c)
≥ 0.
a2 + 2bc
W
Assuming that a ≥ b ≥ c, we get

W
X a(a − b)(a − c) a(a − b)(a − c) b(b − c)(b − a)

≥ +
W
a2 + 2bc a2 + 2bc b2 + 2ca

c(a − b)2 [2a(a − c) + 2b(b − c) + 3a b]
= ≥ 0.
(a2 + 2bc)(b2 + 2ca)

p p p
a3 + b3 + c 3 + 3a bc ≥ a2 a2 + 3bc + b2 b2 + 3ca + c 2 c 2 + 3a b.

Solution. If a = 0, then the inequality is an identity. Consider further that a, b, c > 0,

and write the inequality as follows:
X p
a2 ( a2 + 3bc − a) ≤ 3a bc,
X 3a2 bc
p ≤ 3a bc,
a2 + 3bc + a
X 1
Ç ≤ 1.
1 + 3bc
a 2 + 1
Let us denote
L
1 1 1
x=Ç , y=q , z=Ç ,
.M
3ca
1+ 3bc
a2
+1 1+ b2
+1 1+ 3ab
c2
+1
which implies
D
bc 1 − 2x ca 1 − 2y ab 1 − 2z 1
A
2
= , = , = , 0 < x, y, z < ,
a 3x 2 b 2 3 y2 c 2 3z 2 2
and
PI
(1 − 2x)(1 − 2 y)(1 − 2z) = 27x 2 y 2 z 2 .
M
We need to prove that
LY
x + y +z ≤1
1
O
for 0 < x, y, z < such that (1 − 2x)(1 − 2 y)(1 − 2z) = 27x 2 y 2 z 2 . We will use the
2
1
.M
contradiction method. Assume that x + y + z > 1 for 0 < x, y, z < , and show that
2
(1 − 2x)(1 − 2 y)(1 − 2z) < 27x 2 y 2 z 2 . We have
W
(1 − 2x)(1 − 2 y)(1 − 2z) < (x + y + z − 2x)(x + y + z − 2 y)(x + y + z − 2z)

< ( y + z − x)(z + x − y)(x + y − z)(x + y + z)3
W
≤ 3( y + z − x)(z + x − y)(x + y − z)(x + y + z)(x 2 + y 2 + z 2 )

W
= 3(2x 2 y 2 + 2 y 2 z 2 + 2z 2 x 2 − x 4 − y 4 − z 4 )(x 2 + y 2 + z 2 ).
(2x 2 y 2 + 2 y 2 z 2 + 2z 2 x 2 − x 4 − y 4 − z 4 )(x 2 + y 2 + z 2 ) ≤ 9x 2 y 2 z 2 ,
X
x 6 + y 6 + z 5 + 9x 2 y 2 z 2 ≥ y 2 z 2 ( y 2 + z 2 ).
Clearly, this is just Schur’s inequality of degree three applied to x 2 , y 2 , z 2 . So, the proof
is completed. The equality holds for a = b = c, and also for a = 0 or b = 0 or c = 0.
a b c
p +p +p ≤ 1.
4a2 + 5bc 4b2 + 5ca 4c 2 + 5a b
First Solution (by Vo Quoc Ba Can). If one of a, b, c is zero, then the inequality becomes
an equality. Consider next that a, b, c > 0 and denote
1

a b c
x=p , y=p , z=p , x, y, z ∈ 0, .
4a2 + 5bc 4b2 + 5ca 4c 2 + 5a b 2
L
.M
We have
bc 1 − 4x 2 ca 1 − 4 y2 ab 1 − 4z 2
= , = , = ,
a2 5x 2 b2 5 y2 c2 5z 2
D
and
A
(1 − 4x 2 )(1 − 4 y 2 )(1 − 4z 2 ) = 125x 2 y 2 z 2 .
PI
For the sake of contradiction, assume that x + y + z > 1. Using the AM-GM inequality
M
and the Cauchy-Schwarz inequality, we have
LY
1 Y 1 Y
x 2 y 2z2 = (1 − 4x 2 ) < [(x + y + z)2 − 4x 2 ]
125 125
O
1 Y Y
= (3x + y + z) · ( y + z − x)
125
.M
x + y + z 3 Y
≤ ( y + z − x)
3
1 Y
≤ (x 2 + y 2 + z 2 )(x + y + z) ( y + z − x)
W
9
1
W
= (x 2 + y 2 + z 2 )[2(x 2 y 2 + y 2 z 2 + z 2 x 2 ) − x 4 − y 4 − z 4 ],
9
W
and hence
9x 2 y 2 z 2 < (x 2 + y 2 + z 2 )[2(x 2 y 2 + y 2 z 2 + z 2 x 2 ) − x 4 − y 4 − z 4 ],
X
x 6 + y 6 + z 6 + 3x 2 y 2 z 2 < x 2 y 2 (x 2 + y 2 ).
The last inequality contradicts Schur’s inequality
X
x 6 + y 6 + z 6 + 3x 2 y 2 z 2 ≥ x 2 y 2 (x 2 + y 2 ).
bc ca
Second Solution. In the nontrivial case when a, b, c > 0, setting x = 2
, y = 2 and
a b
ab
z= (x yz = 1), the desired inequality becomes E(x, y, z) ≤ 1, where
c2
1 1 1
E(x, y, z) = p +p +p .
4 + 5x 4 + 5y 4 + 5z
Without loss of generality, we may assume that x ≥ y ≥ z, x ≥ 1, yz ≤ 1. We will prove

that
p p
E(x, y, z) ≤ E(x, yz, yz) ≤ 1.
The left inequality has the form
L
.M
1 1 1
+p ≤p p .
4 + 5y 4 + 5z 4 + 5 yz
p
D
For the nontrivial case y 6= z, consider y > z and denote
A
y +z p
s=
2
, p = yz, PI
M
Æ
q = (4 + 5 y)(4 + 5z).
LY
We have s > p, p ≤ 1 and

Æ Æ
q = 16 + 40s + 25p2 > 16 + 40p + 25p2 = 4 + 5p.
O
.M
By squaring, the desired inequality becomes in succession as follows:

1 1 2 4
+ + ≤ ,
4 + 5 y 4 + 5z q 4 + 5p
W
1 1 2 2 2
W
+ − ≤ − ,
4 + 5 y 4 + 5z 4 + 5p 4 + 5p q
W
8 + 10s 2 2(q − 4 − 5p)

− ≤ ,
q 2 4 + 5p q(4 + 5p)
(s − p)(5p − 4) 8(s − p)
≤ ,
q (4 + 5p)
2 q(4 + 5p)(q + 4 + 5p)
5p − 4 8
≤ ,
q q + 4 + 5p
25p2 − 16 ≤ (12 − 5p)q.
The last inequality is true since
(12 − 5p)q − 25p2 + 16 > (12 − 5p)(4 + 5p) − 25p2 + 16 = 2(8 − 5p)(4 + 5p) > 0.
In order to prove the right inequality, namely

1 2
p +p p ≤ 1,
4 + 5x 4 + 5 yz
p
4 + 5 yz = 3t, t ∈ (2/3, 1]. Since
p
let us denote
1 25
x= = ,
yz (9t − 4)2
2

9t 2 − 4 2
p + ≤ 1,
L
3 36t 4 − 32t 2 + 21 3t
.M
p
(2 − 3t) 36t 4 − 32t 2 + 21 − 3t 2 − 2t ≤ 0.
Since 2 − 3t < 0, we still have to show that
D
A
p
36t 4 − 32t 2 + 21 ≥ 3t 2 + 2t.
Indeed,
PI
M
p 3(t − 1)2 (9t 2 + 14t + 7)
36t 4 − 32t 2 + 21 − 3t 2 − 2t = p ≥ 0.
LY
36t 4 − 32t 2 + 21 + 3t 2 + 2t
O
.M
P 2.32. Let a, b, c be nonnegative real numbers. Prove that

p p p
a 4a2 + 5bc + b 4b2 + 5ca + c 4c 2 + 5a b ≥ (a + b + c)2 .
W
W

W
X p
a 4a2 + 5bc − 2a ≥ 2(a b + bc + ca) − a2 − b2 − c 2 ,
X 1
5a bc p ≥ 2(a b + bc + ca) − a2 − b2 − c 2 .
4a + 5bc + 2a
2
Writing Schur’s inequality

X
a3 + b3 + c 3 + 3a bc ≥ a b(a2 + b2 )
in the form
9a bc
≥ 2(a b + bc + ca) − a2 − b2 − c 2 ,
a+b+c

X 5 9
p ≥ .
4a + 5bc + 2a
2 a+b+c
Let p = a + b + c and q = a b + bc + ca. By the AM-GM inequality, we have

p
p 2 (16a2 + 20bc)(3b + 3c)2 (16a2 + 20bc) + (3b + 3c)2
4a2 + 5bc = ≤
12(b + c) 12(b + c)
16a2 + 16bc + 10(b + c)2 8a2 + 5b2 + 5c 2 + 18bc

≤ = ,
12(b + c) 6(b + c)
L
hence
.M
X 5 X 5
p ≥
4a + 5bc + 2a
2 8a + 5b + 5c 2 + 18bc
2 2
+ 2a
D
6(b + c)
A
X 30(b + c) X 30(b + c)
= = .
8a + 5b + 5c + 12a b + 18bc + 12ac
2 2 2

PI
5p + 2q + 3a2 + 6bc
2
M
X 30(b + c) 9
LY
≥ .
5p2 + 2q + 3a2 + 6bc p
O
30[ (b + c)]2
P
30(b + c)
.M
X
≥P
5p2 + 2q + 3a2 + 6bc (b + c)(5p2 + 2q + 3a2 + 6bc)
W
120p2 120p2
= = .
10p3 + 4pq + 9 bc(b + c) 10p3 + 13pq − 27a bc
P
W

W
120p2 9
≥ ,
10p + 13pq − 27a bc
3 p
10p3 + 81a bc ≥ 39pq.
From Schur’s inequality p3 + 9a bc ≥ 4pq and the known inequality pq ≥ 9a bc, we have
10p3 + 81a bc − 39pq = 10(p3 + 9a bc − 4pq) + pq − 9a bc ≥ 0.
This completes the proof. The equality holds for a = b = c, and also for a = 0 and b = c

X p X a

a 4a + 5bc
2 p ≥ (a + b + c)2 .
4a2 + 5bc
From this inequality and the inequality in P 2.31, namely
X a
p ≤ 1,
4a + 5bc
2
the desired inequality follows.
L
Remark. Using the same way as in the second solution, we can prove the following
inequalities for a, b, c > 0 satisfying a bc = 1:
.M
p p p
a 4a2 + 5 + b 4b2 + 5 + c 4c 2 + 5 ≥ (a + b + c)2 ;
D
p p p
4a4 + 5 + 4b4 + 5 + 4c 4 + 5 ≥ (a + b + c)2 .
A
X
PI
The first inequality follows from the the Cauchy-Schwarz inequality
a
M
X p
a 4a + 5
2 p ≥ (a + b + c)2
4a2 + 5
LY
and the inequality X a

≤ 1, a bc = 1,
O
p
4a2 + 5
.M
which follows from the inequality in P 2.31 by replacing bc/a2 , ca/b2 , a b/c 2 with 1/x 2 ,
1/ y 2 , 1/z 2 , respectively.
The second inequality follows from the the Cauchy-Schwarz inequality
W
2

X p X a
4a4 + 5 p ≥ (a + b + c)2
W
4a + 5
4
W
and the inequality

X a2
p ≤ 1, a bc = 1.
4a4 + 5

p p p
a a2 + 3bc + b b2 + 3ca + c c 2 + 3a b ≥ 2(a b + bc + ca).

First Solution (by Vo Quoc Ba Can). Using the AM-GM inequality yields
X p X a(b + c)(a2 + 3bc)
a a2 + 3bc = p
(b + c)2 (a2 + 3bc)
X 2a(b + c)(a2 + 3bc)
≥ .
(b + c)2 + (a2 + 3bc)
Thus, it suffices to prove the inequality

X 2a(b + c)(a2 + 3bc) X
≥ a(b + c),
a2 + b2 + c 2 + 5bc
L
which can be written as follows
.M
X a(b + c)(a2 − b2 − c 2 + bc)
≥ 0,
D
a2 + b2 + c 2 + 5bc
A
X a3 (b + c) − a(b3 + c 3 )
≥ 0,
a2 + b2 + c 2 + 5bc PI
X a b(a2 − b2 ) − ac(c 2 − a2 )
M
≥ 0,
a2 + b2 + c 2 + 5bc
LY
X a b(a2 − b2 ) X ba(a2 − b2 )
− ≥ 0,
a2 + b2 + c 2 + 5bc b2 + c 2 + a2 + 5ca
O
X 5a bc(a2 − b2 )(a − b)
≥ 0.
(a2 + b2 + c 2 + 5bc)(a2 + b2 + c 2 + 5ac)
.M
Since the last inequality is clearly true, the proof is completed. The equality holds a =
b = c, and also for a = 0 and b = c (or any cyclic permutation).
W

W
X p
(a a2 + 3bc − a2 ) ≥ 2(a b + bc + ca) − a2 − b2 − c 2 .
W
Due to homogeneity, we may assume that a + b + c = 3. By the AM-GM inequality, we

have
p 3a bc 12a bc
a a2 + 3bc − a2 = p = p
a2 + 3bc + a 2 4(a2 + 3bc) + 4a
12a bc
≥ .
4 + a2 + 3bc + 4a
X 1
12a bc ≥ 2(a b + bc + ca) − a2 − b2 − c 2 .
4 + a2 + 3bc + 4a

X 1 9
≥P
4 + a + 3bc + 4a
2 (4 + a + 3bc + 4a)
2
9
=
24 + a + 3 a b
P P
2
27
= P
8( a) + 3 a2 + 9 a b
P P
2
P
9 a 3
= ≥ P .
11( a) + 3 a b
P P
2 4 a
L
Then, it remains to show that
.M
9a bc
≥ 2(a b + bc + ca) − a2 − b2 − c 2 ,
a+b+c
D
A
X X
a3 + 3a bc ≥ a b(a + b).
PI
M
LY

p p p
O
a a2 + 8bc + b b2 + 8ca + c c 2 + 8a b ≤ (a + b + c)2 .

.M
Solution. Multiplying by a + b + c, the inequality becomes

X Æ
a (a + b + c)2 (a2 + 8bc) ≤ (a + b + c)3 .
W
Since
W
Æ
2 (a + b + c)2 (a2 + 8bc) ≤ (a + b + c)2 + (a2 + 8bc),
W

X
a[(a + b + c)2 + (a2 + 8bc)] ≤ 2(a + b + c)3 ,
a3 + b3 + c 3 + 24a bc ≤ (a + b + c)3 .
a(b − c)2 + b(c − a)2 + c(a − b)2 ≥ 0.
a2 + 2bc b2 + 2ca c 2 + 2a b p
p +p +p ≥3 a b + bc + ca.
b2 + bc + c 2 c 2 + ca + a2 a2 + a b + b2
(Michael Rozenberg and Marius Stanean, 2011)
Solution. By the AM-GM inequality, we have

p
X a2 + 2bc X2(a2 + 2bc) a b + bc + ca
p =
b2 + bc + c 2
p
2 (b2 + bc + c 2 )(a b + bc + ca)
L
p X 2(a2 + 2bc)
≥ ab + bc + ca
.M
(b2 + bc + c 2 ) + (a b + bc + ca)
p X 2(a2 + 2bc)
= ab + bc + ca .
(b + c)(a + b + c)
D
A
a2 + 2bc b2 + 2ca c 2 + 2a b
+ +
3
PI
≥ (a + b + c).
b+c c+a a+b
M
2
LY
1X
a4 + b4 + c 4 + a bc(a + b + c) ≥ a b(a + b)2 .
O
2
.M
We can prove this inequality by summing Schur’s inequality of fourth degree

X
a4 + b4 + c 4 + a bc(a + b + c) ≥ a b(a2 + b2 )
W
and the obvious inequality

W
X 1X
a b(a2 + b2 ) ≥ a b(a + b)2 .
W
2
P 2.36. Let a, b, c be nonnegative real numbers, no two of which are zero. If k ≥ 1, then
a k+1 b k+1 c k+1 ak + bk + c k

+ + ≤ .
2a2 + bc 2b2 + ca 2c 2 + a b a+b+c
X ak a k+1

− 2 ≥ 0,
a + b + c 2a + bc
X a k (a − b)(a − c)
≥ 0.
2a2 + bc
Assume that a ≥ b ≥ c. Since (c − a)(c − b) ≥ 0, it suffices to show that
a k (a − b)(a − c) b k (b − a)(b − c)
+ ≥ 0.
L
2a2 + bc 2b2 + ca
.M
This is true if
a k (a − c) b k (b − c)
D
− ≥ 0,
2a2 + bc 2b2 + ca
A
PI
a k (a − c)(2b2 + ca) ≥ b k (b − c)(2a2 + bc).
M
Since a k /b k ≥ a/b, it remains to show that
LY
a(a − c)(2b2 + ca) ≥ b(b − c)(2a2 + bc),

O

.M
(a − b)c[a2 + 3a b + b2 − c(a + b)] ≥ 0.

W
W
tion).
W
a2 − bc b2 − ca c2 − a b
(a) p +p +p ≥ 0;
3a2 + 2bc 3b2 + 2ca 3c 2 + 2a b
a2 − bc b2 − ca c2 − a b
(b) p +p +p ≥ 0.
8a2 + (b + c)2 8b2 + (c + a)2 8c 2 + (a + b)2
Solution. (a) Let

p p p
A= 3a2 + 2bc, B= 3b2 + 2ca, C= 3c 2 + 2a b.
We have
X a2 − bc X (a − b)(a + c) + (a − c)(a + b)
2 =
A A
X (a − b)(a + c) X (b − a)(b + c)
= +
A B
a+c b+c
X
= (a − b) −
A B
L
X a − b (a + c)2 B 2 − (b + c)2 A2
.M
= ·
AB (a + c)B + (b + c)A
X c(a − b)2 2(a − b)2 + c(a + b + 2c)
D
= · ≥ 0.
AB (a + c)B + (b + c)A
A
(b) Let
PI
M
Æ Æ Æ
A= 8a2 + (b + c)2 , B= 8b2 + (c + a)2 , C= 8c 2 + (a + b)2 b.
LY
As we have shown before,

O
X a2 − bc X a − b (a + c)2 B 2 − (b + c)2 A2
2 = · ,
A AB (a + c)B + (b + c)A
.M
hence
X a2 − bc X (a − b)2 C1
2 = · ≥ 0,
W
A AB (a + c)B + (b + c)A
since
W
C1 = [(a + c) + (b + c)][(a + c)2 + (b + c)2 ] − 8ac(b + c) − 8bc(a + c)

W
≥ [(a + c) + (b + c)](4ac + 4bc) − 8ac(b + c) − 8bc(a + c)

= 4c(a − b)2 ≥ 0.
p
P 2.38. Let a, b, c be positive real numbers. If 0 ≤ k ≤ 1 + 2 2, then
a2 − bc b2 − ca c2 − a b
p +p +p ≥ 0.
ka2 + b2 + c 2 k b2 + c 2 + a2 kc 2 + a2 + b2
Solution. Let
p p p
A= ka2 + b2 + c 2 , B= k b2 + c 2 + a2 , C= kc 2 + a2 + b2 .
As we have shown at the preceding problem,

X a2 − bc X a − b (a + c)2 B 2 − (b + c)2 A2
2 = · ,
A AB (a + c)B + (b + c)A
therefore
X a2 − bc X (a − b)2 C1
2 = · ≥ 0,
A AB (a + c)B + (b + c)A
L
where
.M
C1 = (a2 + b2 + c 2 )(a + b + 2c) − (k − 1)c(2a b + bc + ca)
p
D
≥ (a2 + b2 + c 2 )(a + b + 2c) − 2 2 c(2a b + bc + ca).
A
Putting a + b = 2x, we have a2 + b2 ≥ 2x 2 , a b ≤ x 2 , and hence
p
PIp
C1 ≥ (2x 2 + c 2 )(2x + 2c) − 2 2 c(2x 2 + 2c x) = 2(x + c)(x 2 − c)2 ≥ 0.
M
LY
O
.M

p p p
(a2 − bc) b + c + (b2 − ca) c + a + (c 2 − a b) a + b ≥ 0.
W
W
First Solution. Let us denote

v v
W
tb+c s
c+a ta + b
x= , y= , z= ,
2 2 2
hence
a = y 2 + z2 − x 2, b = z2 + x 2 − y 2, c = x 2 + y 2 − z2.
The inequality turns into
x y(x 3 + y 3 ) + yz( y 3 + z 3 ) + z x(z 3 + x 3 ) ≥ x 2 y 2 (x + y) + y 2 z 2 ( y + z) + z 2 x 2 (z + x),
x y(x + y)(x − y)2 + yz( y + z)( y − z)2 + z x(z + x)(z − x)2 ≥ 0.

A(a2 − bc) + B(b2 − ca) + C(c 2 − a b) ≥ 0,
where p p p
A= b + c, B= c + a, C= a + b.
We have
X X
2 A(a2 − bc) = A[(a − b)(a + c) + (a − c)(a + b)]
L
X X
= A(a − b)(a + c) + B(b − a)(b + c)
.M
X
= (a − b)[A(a + c) − B(b + c)]
D
X A2 (a + c)2 − B 2 (b + c)2
= (a − b) ·
A(a + c) + B(b + c)
A
X (a − b)2 (a + c)(b + c)
=
A(a + c) + B(b + c)
PI
≥ 0.
M
LY
O

.M
p p p
(a2 − bc) a2 + 4bc + (b2 − ca) b2 + 4ca + (c 2 − a b) c 2 + 4a b ≥ 0.

W
Solution. If two of a, b, c are zero, then the inequality is clearly true. Otherwise, write
W
the inequality as
AX + BY + C Z ≥ 0,
W
where p p p
a2 + 4bc b2 + 4ca c 2 + 4a b
A= , B= , C= ,
b+c c+a a+b
X = (a2 − bc)(b + c), Y = (b2 − bc)(b + c), X = (c 2 − a b)(a + b).
Without loss of generality, assume that a ≥ b ≥ c. We have X ≥ 0, Z ≤ 0 and
X + Y + Z = 0.
In addition,
X − Y = a b(a − b) + 2(a2 − b2 )c + (a − b)c 2 ≥ 0
and
a4 − b4 + 2(a3 − c 3 )c + (a2 − c 2 )c 2 + 4a bc(a − b) − 4(a − b)c 3

A2 − B 2 =
(b + c)2 (c + a)2
4a bc(a − b) − 4(a − b)c 3 4c(a − b)(a b − c 2 )

≥ = ≥ 0.
(b + c)2 (c + a)2 (b + c)2 (c + a)2
Since
2(AX + BY + C Z) = (A − B)(X − Y ) − (A + B − 2C)Z,
L
A + B − 2C ≥ 0.
.M
This is true if AB ≥ C 2 . Using the Cauchy-Schwarz inequality gives
D
p p
a b + 4c a b a b + 2c a b + 2c 2
A
AB ≥ ≥ .
(b + c)(c + a) (b + c)(c + a)
PI
Thus, we need to show that
M
p
(a + b)2 (a b + 2c a b + 2c 2 ) ≥ (b + c)(c + a)(c 2 + 4a b).
LY

O
p p 2
.M
p
a b(a − b)2 + 2c a b(a + b) a− b + c 2 [2(a + b)2 − 5a b − c(a + b) − c 2 ] ≥ 0.
It is true since
W
2(a + b)2 − 5a b − c(a + b) − c 2 = a(2a − b − c) + b(b − c) + b2 − c 2 ≥ 0.

W
W
tion).

v v v
t a3 t b3 t c3
+ + ≥ 1.
a3 + (b + c)3 b3 + (c + a)3 c 3 + (a + b)3
Solution. For a = 0, the inequality reduces to the obvious inequality

p p p
b3 + c 3 ≥ b3 + c 3 .
For a, b, c > 0, write the inequality as

v
Xu 1
3 ≥ 1.
u
b+c
u
1+
t
a
For any x ≥ 0, we have
L
(1 + x) + (1 − x + x 2 ) 1
.M
p Æ
1 + x3 = (1 + x)(1 − x + x 2 ≤ = 1 + x 2.
2 2
D
Therefore, we get v
1 1
A
Xu X
3 ≥
u
b+c 1 b+c 2
u
t
1+
a
PI
1+
2 a
M
X 1 X a2
≥ = = 1.
b +c 2 2 a2 + b2 + c 2
LY
1+
a2
O
.M

W
v v v
W
u
1 1 1 1 1 1
t t
(a + b + c) + + ≥ 1 + 1 + (a2 + b2 + c 2 ) 2 + 2 + 2 .
t
a b c a b c
W
Solution. Using the Cauchy-Schwarz inequality, we have

v
X X 1 tX X X 1 X 1
a = a2 + 2 bc + 2
a a2 bc
v v
tX X 1 tX X 1

≥ a2 + 2 bc
a2 bc
v v
tX X 1 tX X 1

= a 2 +2 a ,
a2 a
hence
v v
X X 1 tX X 1 tX X 1
a −2 a +1=1+ a2 ,
a a a2
v 2 v
tX X 1 tX X 1
a −1 ≥1+ a2 ,
a a2
v v v
tX X 1 u tX X 1
−1≥ 1+
t
a a2 .
a a2
L
The equality holds if and only if
.M
X X 1 X 1 X
a2 = bc ,
bc a2
D
A
(a2 − bc)(b2 − ca)(c 2 − a b) = 0.
PI
Consequently, the equality occurs for a2 = bc, or b2 = ca, or c 2 = a b.
M
LY

O
v
1 1 1 1 1 1

.M
t
5 + 2(a + b + c ) 2 + 2 + 2 − 2 ≥ (a + b + c)
2 2 2 + + .
a b c a b c

W

W
a b c b c a
x= + + , y= + + .
W
b c a a b c
We have
1 1 1

(a + b + c) + + = x + y +3
a b c
and
1 1 1

2 2 2
2(a + b + c ) 2 + 2 + 2 − 2 =
a b c
2 2 2
2
c 2 a2

a b c b
=2 + + +2 + + +4
b2 c 2 a2 a2 b2 c 2
= 2(x 2 − 2 y) + 2( y 2 − 2x) + 4 = (x + y − 2)2 + (x − y)2 ≥ (x + y − 2)2 .
Therefore, v
1 1 1
t
2(a2+ b2 + c2) + + −2 ≥ x + y −2
a2 b2 c 2
1 1 1

= (a + b + c) + + − 5.
a b c
The equality occurs for a = b, or b = c, or c = a.
L
Æ
2(1 + a bc) + 2(1 + a2 )(1 + b2 )(1 + c 2 ) ≥ (1 + a)(1 + b)(1 + c).
.M
(Wolfgang Berndt, 2006)
D
First Solution. Denoting
A

PI
M
Æ
2(p2 + q2 + r 2 − 2pr − 2q + 1) ≥ p + q − r − 1.
LY
2(p2 + q2 + r 2 − 2pr − 2q + 1) ≥ (p + q − r − 1)2 ,

O
.M
p2 + q2 + r 2 − 2pq + 2qr − 2pr + 2p − 2q − 2r + 1 ≥ 0,

W
(p − q − r + 1)2 ≥ 0.
The equality holds for p + 1 = q + r and q ≥ 1. The last condition follows from p + q −
W
r − 1 ≥ 0.
W
2(1 + a2 ) = (1 + a)2 + (1 − a)2
and
(1 + b2 )(1 + c 2 ) = (b + c)2 + (bc − 1)2 ,
by the Cauchy-Schwarz inequality, we get
Æ
2(1 + a2 )(1 + b2 )(1 + c 2 ) ≥ (1 + a)(b + c) + (1 − a)(bc − 1)
= (1 + a)(1 + b)(1 + c) − 2(1 + a bc).

v v v
t a2 + bc t b2 + ca t c 2 + a b 1
+ + ≥2+ p .
b +c
2 2 c +a
2 2 a +b
2 2
2
Solution. We may assume that a ≥ b ≥ c. Then, it suffices to show that
v v v
t a2 + c 2 t b2 + c 2 t a b 1
+ + ≥2+ p .
b +c
2 2 c +a
2 2 a +b
2 2
2
Let us denote
L
v
t a2 + c 2 s
a
x= , y= .
.M
b2 + c2 b
Since
(a − b)(a b − c 2 )
D
x2 − y2 = ≥ 0,
b(b2 + c 2 )
A
it follows that
x ≥ y ≥ 1. PI
From
M
1 1 (x − y)(x y − 1)

x+ − y+ = ≥ 0,
LY
x y xy
we have v v v
t a2 + c 2 t b2 + c 2
O
s
a tb
+ ≥ + .
b2 + c 2 c 2 + a2 b a
.M

s v v
a t b t ab 1
+ + ≥2+ p .
W
b a a +b
2 2
2
W
s
a
Putting t = , the inequality becomes
b
W
s
1 1 t
t + −2≥ p − .
t 2 t +1
2
We have
(t − 1)2 (t − 1)2
s
1 t
p − = ≤ 2
t +1
2 t +1
s
1

2 t
2(t 2 + 1) p + 2+1
2 t
(t − 1) 2
1
≤ = t + − 2.
t t

Æ Æ Æ Æ
a(2a + b + c) + b(2b + c + a) + c(2c + a + b) ≥ 12(a b + bc + ca).

XÆ
a2 + b2 + c 2 + bc(2b + c + a)(2c + a + b) ≥ 5(a b + bc + ca).
Using the Cauchy-Schwarz inequality yields
L
.M
XÆ XÆ
bc(2b + c + a)(2c + a + b) = (2b2 + bc + a b)(2c 2 + bc + ac)
D
X p X p
≥ (2bc + bc + a bc) = 3(a b + bc + ca) + a bc.
A
X p
PI
a2 + b2 + c 2 + a bc ≥ 2(a b + bc + ca).
M
LY
We can get this inequality by summing Schur’s inequality

X p Xp
a2 + b2 + c 2 + a bc ≥ a b(a + b)
O
.M
and Xp
a b (a + b) ≥ 2(a b + bc + ca).
W
The last inequality is equivalent to the obvious inequality

p
W
Xp p
a b ( a − b )2 ≥ 0.
W
tion).

Æ Æ Æ
a (4a + 5b)(4a + 5c) + b (4b + 5c)(4b + 5a) + c (4c + 5a)(4c + 5b) ≥ 27.

Solution. Assume that a ≥ b ≥ c, consider the non-trivial case b > 0, and write the
inequality in the following equivalent homogeneous forms:
X Æ
a (4a + 5b)(4a + 5c) ≥ 3(a + b + c)2 ,
X X X p p 2
2( a2 − a b) = a 4a + 5b − 4a + 5c ,
X X 25a(b − c)2
(b − c)2 ≥ p p ,
( 4a + 5b + 4a + 5c)2
X
(b − c)2 Sa ≥ 0,
L
where
25a
.M
Sa = 1 − p p .
( 4a + 5b + 4a + 5c)2
Since
D
25b 25b
Sb = 1 − p p ≥1− p p =0
( 4b + 5c + 4b + 5a)2 ( 4b + 9b)2
A
and PI
25c 25c 25
Sc = 1 − p p ≥1− p p =1− > 0,
( 4c + 5a + 4c + 5b)2 ( 9c + 9c)2 36
M
we have
LY
X a2
b
O

a b a
= (b − c)2 Sa + S b .
.M
b a b
b a
Sa + S b ≥ 0.
W
a b
We have
W
p p
25a a( 4a + 5b − 4a)2
Sa ≥ 1 − p p =1− ,
W
( 4a + 5b + 4a)2 b2
p p
25b b( 4b + 5a − 4b)2
Sb ≥ 1 − p p =1− ,
( 4b + 4b + 5a)2 a2
and hence
p p p p
b a b ( 4a + 5b − 4a)2 a ( 4b + 5a − 4b)2
Sa + S b ≥ − + −
a b a b b a
v v
t 4a2 5a t 4b2 5b
a b

=4 + + + − 7 + − 10
b2 b a2 a b a
q p
= 4 4x 2 + 5x − 8 + 2 20x + 41 − 7x − 10,
where
a b
x= + ≥ 2.
b a
To end the proof, we only need to show that x ≥ 2 yields
q p
4 4x 2 + 5x − 8 + 2 20x + 41 ≥ 7x + 10.

p
15x 2 − 60x − 228 + 32 20x + 41 ≥ 0.
L
Indeed,
.M
p p
15x 2 − 60x − 228 + 32 20x + 41 ≥ 15x 2 − 60x − 228 + 32 81 = 15(x − 2)2 ≥ 0.
D
3
The equality holds for a = b = c = 1, and also for c = 0 and a = b = (or any cyclic
A
2
permutation).
PI
M
LY

Æ Æ Æ
(a + 3b)(a + 3c) + b (b + 3c)(b + 3a) + c (c + 3a)(c + 3b) ≥ 12.
O
a
.M
Solution. Assume that a ≥ b ≥ c (b > 0), and write the inequality as

W
X Æ
a (a + 3b)(a + 3c) ≥ 4(a b + bc + ca),
W
X X X p p 2
2( a2 − a b) = a a + 3b − a + 3c ,
W
X X 9a(b − c)2
(b − c)2 ≥ p p ,
( a + 3b + a + 3c)2
X
(b − c)2 Sa ≥ 0,
where
9a
Sa = 1 − p p .
( a + 3b + a + 3c)2
Since
9b 9b
Sb = 1 − p p ≥1− p p =0
( b + 3c + b + 3a)2 ( b + 4b)2
and
9c 9c 9
Sc = 1 − p p ≥1− p p =1− > 0,
( c + 3a + c + 3b)2 ( 4c + 4c)2 16
we have
X a2
b

a b a
= (b − c)2 Sa + S b .
b a b
b a
L
Sa + S b ≥ 0.
a b
.M
We have p p
9a a( a + 3b − a)2
Sa ≥ 1 − p =1− ,
D
p
( a + 3b + a)2 b2
p p
A
9b b( b + 3a − b)2
Sb ≥ 1 − p p =1− ,
( b + b + 3a)2 a2 PI
and hence
M
p p p p
b a b ( a + 3b − a)2 a ( b + 3a − b)2
LY
Sa + S b ≥ − + −
a b a b b a
v v
t a2 3a t b2 3b
O
a b
=2 + + + − + −6
b2 b a2 a b a
.M
Æ p
= 2 x 2 + 3x − 2 + 2 3x + 10 − x − 6,
where
W
a b
x= + ≥ 2.
b a
W
To end the proof, it remains to show that

W
Æ p
2 x 2 + 35x − 2 + 2 3x + 10 ≥ x + 6
for x ≥ 2. By squaring, this inequality becomes

p
3x 2 − 44 + 8 3x + 10 ≥ 0.
Indeed, p
3x 2 − 44 + 8 3x + 10 ≥ 12 − 44 + 32 = 0.
p
The equality holds for a = b = c = 1, and also for c = 0 and a = b = 3 (or any cyclic
permutation).

p p p Æ
2 + 7a b + 2 + 7bc + 2 + 7ca ≥ 3 3(a b + bc + ca).
Solution. Consider a ≥ b ≥ c. Since the inequality is trivial for b = c = 0, we may
assume that b > 0. By squaring, the desired inequality becomes
XÆ
a2 + b2 + c 2 + (2 + 7a b)(2 + 7ac) ≥ 10(a b + bc + ca),
X p p 2
6(a2 + b2 + c 2 − a b − bc − ca) = 2 + 7a b − 2 + 7ac ,
L
X X 49a2 (b − c)2
3 (b − c)2 ≥ p p ,
.M
( 2 + 7a b + 2 + 7ac )2
X
(b − c)2 Sa ≥ 0,
D
where
A
49a2
Sa = 1 − p p ,
( 6 + 21a b + 6 + 21ac)2
PI
49b2
Sb = 1 − p
M
p ,
( 6 + 21a b + 6 + 21bc)2
LY
49c 2
Sc = 1 − p p .
( 6 + 21ac + 6 + 21bc)2
O
Since 6 ≥ 2(a2 + b2 ) ≥ 4a b, we have

.M
49a2 49a2 a
Sa ≥ 1 − p p ≥1− p p =1− ,
( 4a b + 21a b + 4a b + 21ac)2 (5 a b + 2 a b)2 b
W
49b2 49b2 b
Sb ≥ 1 − p p ≥1− p p =1− ,
( 4a b + 21a b + 4a b + 21bc)2 (5 a b + 2 a b)2 a
W
49c 2 49c 2 49
Sc ≥ 1 − p p ≥1− =1− > 0.
W
( 4a b + 21ac + 4a b + 21bc) 2 (5c + 5c)2 100

Therefore,
X
(b − c)2 Sa ≥ (b − c)2 Sa + (c − a)2 S b

2
a 2 b
≥ (b − c) 1 − + (c − a) 1 −
b a
(a − b)2 (a b − c 2 )
= ≥ 0.
ab
p
The equality holds for a = b = c = 1, and also for c = 0 and a = b = 3 (or any cyclic
permutation).

Pp
(a) a(b + c)(a2 + bc) ≥ 6;
p p
a(b + c) a2 + 2bc ≥ 6 3;
P
(b)
p
a(b + c) (a + 2b)(a + 2c) ≥ 18.
P
(c)
Solution. Assume that a ≥ b ≥ c (b > 0).

(a) Write the inequality in the homogeneous form
L
.M
XÆ
a(b + c)(a2 + bc) ≥ 2(a b + bc + ca).
D
First Solution. Write the homogeneous inequality as
A
XÆ p Æ
a(b + c) a2 + bc − a(b + c) ≥ 0,
p
X (a − b)(a − c) a(b + c)
PI
M
p p ≥ 0.
a2 + bc + a(b + c)
LY
Since (c − a)(c − b) ≥ 0, it suffices to show that

O
p p
(a − b)(a − c) a(b + c) (b − c)(b − a) b(c + a)
p p + p p ≥ 0.
a2 + bc + a(b + c) b2 + ca + b(c + a)
.M
This is true if p p
(a − c) a(b + c) (b − c) b(c + a)
W
p p ≥p p .
a2 + bc + a(b + c) b2 + ca + b(c + a)
W
Since Æ Æ
a(b + c) ≥ b(c + a),
W

a−c b−c
p p ≥p p .
a2 + bc + a(b + c) b2 + ca + b(c + a)
Moreover, since
p Æ p Æ
a2 + bc ≥ a(b + c), b2 + ca ≤ b(c + a),

a−c b−c
p ≥p .
a2 + bc b2 + ca
Indeed, we have
(a − c)2 (b2 + ca) − (b − c)2 (a2 + bc) = (a − b)(a2 + b2 + c 2 + 3a b − 3bc − 3ca) ≥ 0,
because
a2 + b2 + c 2 + 3a b − 3bc − 3ca = (a2 − bc) + (b − c)2 + 3a(b − c) ≥ 0.

p
The equality holds for a = b = c = 1, and for c = 0 and a = b = 3 (or any cyclic
permutation).
Second Solution. By squaring, the homogeneous inequality becomes
L
.M
X XÆ
a(b + c)(a2 + bc) + 2 bc(a + b)(a + c)(b2 + ca)(c 2 + a b) ≥ 4(a b + bc + ca)2 .
Since
D
(b2 + ca)(c 2 + a b) − bc(a + b)(a + c) = a(b + c)(b − c)2 ≥ 0,
A
X X
PI
a(b + c)(a2 + bc) + 2 bc(a + b)(a + c) ≥ 4(a b + bc + ca)2 ,
M
LY
which is equivalent to X
bc(b − c)2 ≥ 0.
O
(b) Write the inequality as

.M
X p p
a(b + c) a2 + 2bc ≥ 2(a b + bc + ca) a b + bc + ca,
W
X p p
a(b + c) a2 + 2bc − a b + bc + ca ≥ 0,
W
X a(b + c)(a − b)(a − c)

p p ≥ 0.
a2 + 2bc + a b + bc + ca
W
a(b + c)(a − b)(a − c) b(c + a)(b − c)(b − a)

p p +p p ≥ 0.
a + 2bc + a b + bc + ca
2 b2 + 2ca + a b + bc + ca
This is true if
a(b + c)(a − c) b(c + a)(b − c)
p p ≥p p .
a2 + 2bc + a b + bc + ca b2 + 2ca + a b + bc + ca
Since
(b + c)(a − c) ≥ (c + a)(b − c),

a b
p p ≥p p .
a2 + 2bc + a b + bc + ca b2 + 2ca + a b + bc + ca
Moreover, since
p p p p
a2 + 2bc ≥ a b + bc + ca, b2 + 2ca ≤ a b + bc + ca,

a b
p ≥p .
a2 + 2bc b2 + 2ca
L
Indeed, we have
.M
a2 (b2 + 2ca) − b2 (a2 + 2bc) = 2c(a3 − b3 ) ≥ 0.
p
The equality holds for a = b = c = 1, and for c = 0 and a = b =
D
3 (or any cyclic
permutation).
A
(c) Write the inequality as follows:
X Æ Æ
PI
a(b + c) (a + 2b)(a + 2c) ≥ 2(a b + bc + ca) 3(a b + bc + ca),
M
X Æ Æ
LY
a(b + c) (a + 2b)(a + 2c) − 3(a b + bc + ca) ≥ 0,

X a(b + c)(a − b)(a − c)
O
p p ≥ 0.
(a + 2b)(a + 2c) + 3(a b + bc + ca)
.M

a(b + c)(a − b)(a − c) b(c + a)(b − c)(b − a)
+p ≥ 0.
W
p p p
(a + 2b)(a + 2c) + 3(a b + bc + ca) (b + 2c)(b + 2a) + 3(a b + bc + ca)
W
This is true if
a(b + c)(a − c) b(c + a)(b − c)
W
p p ≥p p .
Since
(b + c)(a − c) ≥ (c + a)(b − c),
a b
p p ≥p p .
Moreover, since
Æ Æ Æ Æ
(a + 2b)(a + 2c) ≥ 3(a b + bc + ca), (b + 2c)(b + 2a) ≤ 3(a b + bc + ca),
a b
p ≥p .
(a + 2b)(a + 2c) (b + 2c)(b + 2a)
This is true if p p
a b
p ≥p .
(a + 2b)(a + 2c) (b + 2c)(b + 2a)
Indeed, we have
a(b + 2c)(b + 2a) − b(a + 2b)(a + 2c) = (a − b)(a b + 4bc + 4ca) ≥ 0.
L
p
.M
The equality holds for a = b = c = 1, and for c = 0 and a = b = 3 (or any cyclic
permutation).
D
A
PI
p
M
p p
a bc + 3 + b ca + 3 + c a b + 3 ≥ 6.
LY

O
First Solution. Write the inequality in the homogeneous form

.M
X p
a a b + 2bc + ca ≥ 2(a b + bc + ca).
If a = 0, then the inequality turns into

W
p p p
bc( b − c)2 ≥ 0.
W
Consider further a, b, c > 0. By squaring, the inequality becomes

W
X X Æ X X
a b(a2 + b2 ) + 2 bc (bc + 2ca + a b)(ca + 2a b + bc) ≥ 4 a2 b2 + 6a bc a.

Æ Æ
(bc + 2ca + a b)(ca + 2a b + bc) = (a b + bc + ca + ca)(a b + bc + ca + a b)
p
≥ a b + bc + ca + a bc,
and hence
X Æ Xp
bc (bc + 2ca + a b)(ca + 2a b + bc) ≥ (a b + bc + ca)2 + a bc bc.

X Xp X X
a b(a2 + b2 ) + 2(a b + bc + ca)2 + 2a bc bc ≥ 4 a2 b2 + 6a bc a,
X Xp X X
a b(a2 + b2 ) + 2a bc bc ≥ 2 a2 b2 + 2a bc a,
X X Xp
a b(a − b)2 ≥ 2a bc( a− bc),
X (a − b)2 Xp p
≥ ( a − b)2 .
L
c
.M
p p p
Using the substitution x = a, y = b, z = c, the last inequality becomes
D
X (x 2 − y 2 )2 X
≥ (x − y)2 ,
z2
A
PI
( y − z)2 A + (z − x)2 B + (x − y)2 C ≥ 0,
where
M
y +z− x z+x− y x + y −z
A= , B= , C= .
x2 y2 z2
LY
Without loss of generality, assume that x ≥ y ≥ z > 0. Since B > 0 and C > 0, we have
O
( y − z)2 A + (z − x)2 B + (x − y)2 C ≥ ( y − z)2 A + (z − x)2 B ≥ ( y − z)2 (A + B).

.M
Thus, we only need to show that A + B ≥ 0. Indeed,
y +z− x z+ x − y y +z− x z+ x − y
W
2z
A+ B = + ≥ + = 2 > 0.
x2 y2 x2 x2 x
W
p
W
permutation).
Second Solution. Assume that a ≥ b ≥ c, and write the inequality as
X p
a a b + 2bc + ca ≥ 2(a b + bc + ca),
X p
a a b + 2bc + ca − b − c ≥ 0,
X a(a b + ac − b2 − c 2 )
p ≥ 0,
a b + 2bc + ca + b + c
x y z
+ + ≥ 0, (*)
b+c+A c+a+B a+ b+C
where
p p p
A= a b + 2bc + ca, B= bc + 2ca + a b, C= ca + 2a b + bc,
x = a2 (b + c) − a(b2 + c 2 ), y = b2 (c + a) − b(c 2 + a2 ), z = c 2 (a + b) − c(a2 + b2 ).

Since x + y + z = 0, we can write the inequality as
1 1 1 1

x − +z − ≥ 0.
b+c+A c+a+B a+b+C c+a+B
We have
x = a b(a − b) + ac(a − c) ≥ 0, z = ac(c − a) + bc(c − b) ≤ 0.
L
.M
1 1 1 1
− ≥ 0, − ≤ 0,
D
b+c+A c+a+B a+b+C c+a+B
A
that is,
a − b + B − A ≥ 0, b − c + C − B ≥ 0.
PI
It is enough to prove that A ≤ B ≤ C. Indeed,
M
B 2 − A2 = c(a − b) ≥ 0, C 2 − B 2 = a(b − c) ≥ 0.
LY
Remark. We can also prove the inequality (*) as follows

O
X a b(a − b) + ac(a − c)
≥ 0,
b+c+A
.M
+ ≥ 0,
b+c+A c+a+B
W
1 1
X
a b(a − b) − ≥ 0,
b+c+A c+a+B
W
X
a b(a + b + C)(a − b)(a − b + B − A) ≥ 0,
W
X
2
c
a b(a + b + C)(a − b) 1 + ≥ 0.
A+ B

p
(b + c) b2 + c 2 + 7bc ≥ 18;
P
(a)
p p
(b + c) b2 + c 2 + 10bc ≤ 12 3.
P
(b)
Solution. (a) Write the inequality in the equivalent homogeneous forms

X p
(b + c) b2 + c 2 + 7bc ≥ 2(a + b + c)2 ,
X p
(b + c) b2 + c 2 + 7bc − b2 − c 2 − 4bc ≥ 0,
X (b + c)2 (b2 + c 2 + 7bc) − (b2 + c 2 + 4bc)2
p ≥ 0,
(b + c) b2 + c 2 + 7bc + b2 + c 2 + 4bc
X bc(b − c)2
p ≥ 0.
(b + c) b2 + c 2 + 7bc + b2 + c 2 + 4bc
L
3
The equality holds for a = b = c = 1, for a = 0 and b = c = (or any cyclic permuta-
.M
2
tion), and for a = 3 and b = c = 0 (or any cyclic permutation).
D
(b) Write the inequality as
A
X Æ
(b + c) 3(b2 + c 2 + 10bc) ≤ 4(a + b + c)2 ,
X Æ
PI
2b2 + 2c 2 + 8bc − (b + c) 3(b2 + c 2 + 10bc) ≥ 0,
M
X 4(b2 + c 2 + 4bc)2 − 3(b + c)2 (b2 + c 2 + 10bc)
LY
p ≥ 0,
2b2 + 2c 2 + 8bc + (b + c) 3(b2 + c 2 + 10bc)
(b − c)4
O
X
p ≥ 0.
2b2 + 2c 2 + 8bc + (b + c) 3(b2 + c 2 + 10bc)
.M

W
W
P 2.53. Let a, b, c be nonnegative real numbers such then a + b + c = 2. Prove that

W
p p p p
a + 4bc + b + 4ca + c + 4a b ≥ 4 a b + bc + ca.
Solution. Without loss of generality, assume that c = min{a, b, c}. Using Minkowski’s
inequality gives
p p p
p p Ç p p p p
a + 4bc + b + 4ca ≥ ( a + b)2 + 4c( a + b)2 = ( a + b) 1 + 4c.

p p p p p
( a + b) 1 + 4c ≥ 4 a b + bc + ca − c + 4a b.

p Æ
(a + b + 2 a b)(1 + 4c) + 8 (a b + bc + ca)(c + 4a b) ≥ 16(a b + bc + ca) + c + 4a b.
According to Lemma below, it suffices to show that
p
(a + b + 2 a b)(1 + 4c) + 8(2a b + bc + ca) ≥ 16(a b + bc + ca) + c + 4a b,
p p
a+ b−c+2 a b + 8c a b ≥ 4(a b + bc + ca).
Write this inequality in the homogeneous form
L
p p
(a + b + c)(a + b − c + 2 a b) + 16c a b ≥ 8(a b + bc + ca).
.M
p 1
Due to homogeneity, we may assume that a + b = 1. Let us denote d = a b, 0 ≤ d ≤ .
2
D
We need to show that f (c) ≥ 0 for 0 ≤ c ≤ d, where
A
f (c) = (1 + c)(1 − c + 2d) + 16cd − 8d 2 − 8c
PI
= (1 − 2d)(1 + 4d) + 2(9d − 4)c − c 2 .
M
Since f (c) is concave, it suffices to show that f (0) ≥ 0 and f (d) ≥ 0. Indeed,
f (0) = (1 − 2d)(1 + 4d) ≥ 0,
LY
f (d) = (3d − 1)2 ≥ 0.

O
Thus, the proof is completed. The equality holds for a = b = 1 and c = 0 (or any cyclic
permutation).
.M
Lemma (by Nguyen Van Quy). Let a, b, c be nonnegative real numbers such then
c = min{a, b, c}, a + b + c = 2.
W
Then,
W
Æ
(a b + bc + ca)(c + 4a b) ≥ 2a b + bc + ca.
W
Proof. By squaring, the inequality becomes

c[a b + bc + ca − c(a + b)2 ] ≥ 0.
(a + b + c)(a b + bc + ca) − 2c(a + b)2 ≥ 0.
We have
(a + b + c)(a b + bc + ca) − 2c(a + b)2 ≥ (a + b)(b + c)(c + a) − 2c(a + b)2
= (a + b)(a − c)(b − c) ≥ 0.

p p p p
a2 + b2 + 7a b + b2 + c 2 + 7bc + c 2 + a2 + 7ca ≥ 5 a b + bc + ca.
Solution (by Nguyen Van Quy). Assume that c = min{a, b, c}. Using Minkowski’s in-
equality yields
p
p p Ç p
b + c + 7bc + a + c + 7ca ≥ (a + b)2 + 4c 2 + 7c( a + b)2 .
2 2 2 2
L
.M
p
Ç p p p
(a + b)2 + 4c 2 + 7c( a + b)2 ≥ 5 a b + bc + ca − a2 + b2 + 7a b.
D
A
p Æ
2c 2 + 7c a b + 5 (a2 + b2 + 7a b)(a b + bc + ca) ≥ 15a b + 9c(a + b).
PI
Due to homogeneity, we may assume that a + b = 1. Let us denote x = a b. We need to
1
M
show that f (x) ≥ 0 for c 2 ≤ x ≤ , where
4
LY
p Æ
f (x) = 2c 2 + 7c x + 5 (1 + 5x)(c + x) − 15x − 9c.
O
Since
−7c 5(5c − 1)2
f 00 (x) = p − p <0
.M
4 x 3 4 [5x 2 + (5c + 1)x + c]3

1

f (c) is concave. Thus, it suffices to show that f (c 2 ) ≥ 0 and f ≥ 0.
W
4
Write the inequality f (c 2 ) ≥ 0 as
W
Æ
5 (1 + 5c 2 )(c + c 2 ) ≥ 6c 2 + 9c.
W
By squaring, this inequality turns into
c(89c 3 + 17c 2 − 56c + 25) ≥ 0,
which is true since
89c 3 + 17c 2 − 56c + 25 ≥ 12c 2 − 56c + 25 = (1 − 2c)(25 − 6c) ≥ 0.

1

Write the inequality f ≥ 0 as
4
p
8c 2 − 22c + 15( 4c + 1 − 1) ≥ 0.
p
Making the substitution t = 4c + 1, t ≥ 1, the inequality becomes
(t − 1)(t 3 + t 2 − 12t + 18) ≥ 0.
This is true since
t 3 + t 2 − 12t + 18 ≥ 2t 2 − 12t + 18 = 2(t − 3)2 ≥ 0.
Thus, the proof is completed. The equality holds for a = b and c = 0 (or any cyclic
permutation).
L
.M
D
p p p Æ
a2 + b2 + 5a b + b2 + c 2 + 5bc + c 2 + a2 + 5ca ≥ 21(a b + bc + ca).
A
PI (Nguyen Van Quy, 2012)
Solution. Without loss of generality, assume that c = min{a, b, c}. Using Minkowski’s
M
inequality, we have
LY
p
Æ Æ Ç p
(a + c)2 + 3ac + (b + c)2 + 3bc ≥ (a + b + 2c)2 + 3c( a + b)2 .
O

.M
p
Ç p Æ p
(a + b + 2c)2 + 3c( a + b)2 ≥ 21(a b + bc + ca) − a2 + b2 + 5a b.
W

W
p Æ
2c 2 + 3c ab + 21(a2 + b2 + 5a b)(a b + bc + ca) ≥ 12a b + 7c(a + b).
W
Due to homogeneity, we may assume that a + b = 1. Let us denote x = a b. We need to

1
show that f (x) ≥ 0 for c 2 ≤ x ≤ , where
4
p Æ
f (x) = 2c 2 + 3c x + 21(1 + 3x)(c + x) − 12x − 7c.
Since p
00 −3c 21(3c − 1)2
f (x) = p − p <0
4 x 3 4 [3x 2 + (3c + 1)x + c]3
1

f (c) is concave. Thus, it suffices to show that f (c 2 ) ≥ 0 and f ≥ 0.
4
Write the inequality f (c 2 ) ≥ 0 as

Æ
21(1 + 3c 2 )(c + c 2 ) ≥ 7(c + c 2 ).
By squaring, this inequality turns into
c(c + 1)(1 − 2c)(3 − c) ≥ 0,
which is clearly true.

1
Write the inequality f ≥ 0 as
4
L
Æ
8c 2 − 22c + 7 3(4c + 1) − 12 ≥ 0.
.M
1
Using the substitution 3t 2 = 4c + 1, t ≥ p , the inequality becomes
D
3
A
(t − 1)2 (3t 2 + 6t − 4) ≥ 0.
This is true since

PI
p
M
3t 2 + 6t − 4 ≥ 1 + 2 3 − 4 > 0.
LY
Thus, the proof is completed. The equality holds for a = b = c.

O
.M

W
v
p p p t2
a a2 + 5 + b b2 + 5 + c c2 + 5 ≥ (a + b + c)2 .
3
W

W

X Æ p
a 3a2 + 5(a b + bc + ca) ≥ 2 (a + b + c)2 .
Due to homogeneity, we may assume that
a b + bc + ca = 1.

X X Æ X X X
a4 + 2 bc (3b2 + 5)(3c 2 + 5) ≥ 12 a2 b2 + 19a bc a+3 a b(a2 + b2 ).
Applying Lemma below for x = 3b2 , y = 3c 2 and d = 5, we have

Æ 9 2
2 (3b2 + 5)(3c 2 + 5) ≥ 3(b2 + c 2 ) + 10 − (b − c 2 )2 ,
20
hence
Æ 9
2bc (3b2 + 5)(3c 2 + 5) ≥ 3bc(b2 + c 2 ) + 10bc − bc(b2 − c 2 )2 ,
20
X Æ X X 9 X
2 bc (3b2 + 5)(3c 2 + 5) ≥ 3 bc(b2 + c 2 ) + 10( bc)2 − bc(b2 − c 2 )2
20
X X X 9 X
L
= 10 a2 b2 + 20a bc a+3 a b(a2 + b2 ) − bc(b2 − c 2 )2 .
20
.M
9 X
D
X X X X
a4 + 10 a2 b2 + 20a bc a+3 a b(a2 + b2 ) − bc(b2 − c 2 )2 ≥
20
A
X X X
≥ 12 a2 b2 + 19a bc a+3 a b(a2 + b2 ),
PI
M
X X X 9 X
a4 − 2 a2 b2 + a bc a− bc(b2 − c 2 )2 ≥ 0.
LY
20
Since
O
X X X X X X X
2 a4 − 2 a2 b2 + a bc a =2 a4 − a2 b2 − 2 a2 b2 − a bc a
.M
X X
= (b2 − c 2 )2 − a2 (b − c)2 ,
W

W
X
(b − c)2 Sa ≥ 0,
W
where
9
Sa = (b + c)2 − a2 − bc(b + c)2 .
10
In addition, since
bc(b + c)2
Sa ≥ (b + c)2 − a2 − bc(b + c)2 = (b + c)2 − a2 − ,
a b + bc + ca
a(b + c)3 − a2 (a b + bc + ca)
= ,
a b + bc + ca
it is enough to show that X
(b − c)2 Ea ≥ 0,
where
Ea = a(b + c)3 − a2 (a b + bc + ca).
Assume that a ≥ b ≥ c, b > 0. Since
E b = b(c + a)3 − b2 (a b + bc + ca) ≥ b(c + a)3 − b2 (c + a)(c + b)
≥ b(c + a)3 − b2 (c + a)2 = b(c + a)2 (c + a − b) ≥ 0,
Ec = c(a + b)3 − c 2 (a b + bc + ca) ≥ c(a + b)3 − c 2 (a + b)(b + c)
≥ c(a + b)3 − c 2 (a + b)2 = c(a + b)2 (a + b − c) ≥ 0
L
.M
and
Ea E b (b + c)3 (c + a)3
+ = + − 2(a b + bc + ca)
D
a2 b2 a b
b3 + 2b2 c a3 + 2a2 c
A
≥ + − 2(a b + bc + ca)
=
a b PI
(a2 − b2 )2 + 2c(a + b)(a − b)2
≥ 0,
M
ab
we get
LY
2 Ea Eb
X
2 2 2 2
(b − c) Ea ≥ (b − c) Ea + (a − c) E b ≥ a (b − c) +
O
≥ 0.
a2 b2
.M
p
The equality holds for a = b = c = 1, and also for a = b = 3 and c = 0 (or any cyclic
permutation).
W
Lemma. If x ≥ 0, y ≥ 0 and d > 0, then

W
Æ 1
2 (x + d)( y + d) ≥ x + y + 2d − (x − y)2 .
4d
W
Proof. We have
Æ 2(d x + d y + x y) 2(d x + d y + x y)
2 (x + d)( y + d) − 2d = p ≥ (x+d)+( y+d)
(x + d)( y + d) + d 2 +d
4(d x + d y + x y) (x − y)2 (x − y)2

= =x+y− ≥x+y− .
x + y + 4d x + y + 4d 4d

p p p
a 2 + 3bc + b 2 + 3ca + c 2 + 3a b ≥ (a + b + c)2 .
Solution. Let q = a b + bc + ca. Write the inequality as

X p
a 2 + 3bc ≥ 1 + 2q.
L
X X Æ
1 + 3a bc a+2 bc (2 + 3a b)(2 + 3ac) ≥ 4q + 4q2 .
.M
Applying Lemma from the preceding P 2.56 for x = 3a b, y = 3ac2 and d = 2, we have
D
Æ 9
2 (2 + 3a b)(2 + 3ac) ≥ 3a(b + c) + 4 − a2 (b − c)2 ,
A
8
2
X Æ
bc (2 + 3a b)(2 + 3ac) ≥ 3a bc
X
(b + c) + 4
X PI 9
bc − a bc
X
a(b − c)2
8
M
X 9 X
= 6a bc a + 4q − a bc a(b − c)2 .
LY
8
O
X X 9 X
1 + 3a bc a + 4q + 6a bc a − a bc a(b − c)2 ≥ 4q + 4q2 ,
.M
8
W
X 9 X
1 + 9a bc a − 4q2 ≥ a bc a(b − c)2 .
8
W
Since
W
X
a4 + b4 + c 4 = 1 − 2(a2 b2 + b2 c 2 + c 2 a2 ) = 1 − 2q2 + 4a bc a,
from Schur’s inequality of fourth degree
X X X
a4 + b4 + c 4 + 2a bc a≥ a2 ab ,
we get X
1 ≥ 2q2 + q − 6a bc a.
X X 9 X
(2q2 + q − 6a bc a) + 9a bc a − 4q2 ≥ a bc a(b − c)2 ;
8
that is, X X
8(q − 2q2 + 3a bc a) ≥ 9a bc a(b − c)2 .
Since X X X X 2 X
q − 2q2 + 3a bc a= a2 ab − 2 a b + 3a bc a
X X X
= bc(b2 + c 2 ) − 2 b2 c 2 = bc(b − c)2 ,
we need to show that X

bc(8 − 9a2 )(b − c)2 ≥ 0.
Also, since
L
8 − 9a2 = 8(b2 + c 2 ) − a2 ≥ b2 + c 2 − a2 ,
.M
D
X
bc(b2 + c 2 − a2 )(b − c)2 ≥ 0.
A
Assume that a ≥ b ≥ c. It is enough to show that PI
bc(b2 + c 2 − a2 )(b − c)2 + ca(c 2 + a2 − b2 )(c − a)2 ≥ 0.
M
LY
This is true if
a(c 2 + a2 − b2 )(a − c)2 ≥ b(a2 − b2 − c 2 )(b − c)2 .
O
For the non-trivial case a2 − b2 − c 2 ≥ 0, this inequality follows from

.M
a ≥ b, c 2 + a2 − b2 ≥ a2 − b2 − c 2 , (a − c)2 ≥ (b − c)2 .
1 1
W
The equality holds for a = b = c = p , and for a = 0 and b = c = p (or any cyclic
3 2
permutation).
W
W

v v v
t 2a + bc t 2b + ca t 2c + a b
(a) a +b +c ≥ 3;
3 3 3
v v v
t a(1 + b + c) t b(1 + c + a) t c(1 + a + b)
(b) a +b +c ≥ 3.
3 3 3
Solution. (a) If two of a, b, c are zero, then the inequality is trivial. Otherwise, by
Hölder’s inequality, we have
X t 2a + bc 2
v
( a)3
P
9
a ≥P =P a .
3 3a
2a + bc 2a + bc

X a
≤ 1.
2a + bc
L
Since
.M
2a bc
=1− ,
2a + bc 2a + bc
D
we can write this inequality as
X bc
A
≥ 1.
2a + bc
By the Cauchy-Schwarz inequality, we have PI
M
( bc)2 ( bc)2
P P
X bc
≥P = = 1.
2a + bc bc(2a + bc) 2a bc a + b2 c 2
P P
LY
3
O
The equality holds for a = b = c = 1, and for a = 0 and b = c = (or any cyclic
2
permutation).
.M
(b) Write the inequality in the homogeneous form

W
X Æ
a a(a + 4b + 4c) ≥ (a + b + c)2 .
W

W
X Æ X X
bc bc(b + 4c + 4a)(c + 4a + 4b) ≥ 3 a2 b2 + 6a bc a.
Applying the Cauchy-Schwarz inequality, we have

Æ Æ
(b + 4c + 4a)(c + 4a + 4b) = (4a + b + c + 3c)(4a + b + c + 3b)
p
≥ 4a + b + c + 3 bc,
hence Æ p
bc bc(b + 4c + 4a)(c + 4a + 4b) ≥ (4a + b + c)bc bc + 3b2 c 2 ,
X Æ X p X
bc bc(b + 4c + 4a)(c + 4a + 4b) ≥ (4a + b + c)bc bc + 3 b2 c 2 .

X p X
(4a + b + c)bc bc ≥ 6a bc a.
Replacing a, b, c by a2 , b2 , c 2 , respectively, this inequality becomes

X X
(4a2 + b2 + c 2 )b3 c 3 ≥ 6a2 b2 c 2 a2 ,
X X X X
a2 b3 c 3 + 3a2 b2 c 2 bc ≥ 6a2 b2 c 2 a2 ,
X X X X
a2 a3 b3 − 3a2 b2 c 2 ≥ 3a2 b2 c 2 a2 − ab .
L
Since
.M
X X X X 1 X X
a3 b3 − 3a2 b2 c 2 = ab a2 b2 − a bc a = ab a2 (b − c)2 ,
2
D
and X X 1X
A
a2 − ab = (b − c)2 ,
2
we can write the inequality as PI
M
X
(b − c)2 Sa ≥ 0,
LY
where X X
Sa = a 2 a2 a b − 3a2 b2 c 2 .
O
Assume that a ≥ b ≥ c. Since Sa ≥ S b ≥ 0 and

.M
X X
S b + Sc = (b2 + c 2 ) a2 a b − 6a2 b2 c 2
X X
≥ 2bc a2 a b − 6a2 b2 c 2
W
X
≥ 2bca2 a b − 6a2 b2 c 2 = 2a2 bc(a b + ac − 2bc) ≥ 0,
W
we get X
W
(b − c)2 Sa ≥ (c − a)2 S b + (a − b)2 Sc ≥ (a − b)2 (S b + Sc ) ≥ 0.

3
The equality holds for a = b = c = 1, and for a = 0 and b = c = (or any cyclic
2
permutation).

Æ Æ Æ
8(a2 + bc) + 9 + 8(b2 + ca) + 9 + 8(c 2 + a b) + 9 ≥ 15.

Solution. Let q = a b + bc + ca and
A = (3a − b − c)2 + 8q, B = (3b − c − a)2 + 8q, C = (3c − a − b)2 + 8q.
Since
8(a2 + bc) + 9 = 8(a2 + q) + 9 − 8a(b + c) = 8(a2 + q) + 9 − 8a(3 − a)

= (4a − 3)2 + 8q = (3a − b − c)2 + 8q = A,
we can rewrite the inequality as follows

Xp
A ≥ 15,
L
X p
.M
[ A − (3a + b + c)] ≥ 0,
X 2bc − ca − a b
D
p ≥ 0,
A + 3a + b + c
A
X b(c − a) c(b − a)

p + p ≥ 0,
A + 3a + b + c PI
A + 3a + b + c
X c(a − b) X c(b − a)
M
p + p ≥ 0,
B + 3b + c + a A + 3a + b + c
p p p
LY
X
c(a − b)( C + 3c + a + b)[ A − B + 2(a − b)] ≥ 0,
p 4(a + b − c)
O
X
c(a − b)2 ( C + 3c + a + b) p p + 1 ≥ 0.
A+ B
.M
Without loss of generality, assume that a ≥ b ≥ c. Since a + b − c > 0, it suffices to show

that
W
p 4(c + a − b)

2
b(a − c) ( B + 3b + c + a) p p +1 ≥
A+ C
W
p 4(a − b − c)

a(b − c)2 ( A + 3a + b + c) p p − 1 .
B+ C
W
This inequality follows from the inequalities
b2 (a − c)2 ≥ a2 (b − c)2 ,
p p
a( B + 3b + c + a) ≥ b( A + 3a + b + c),
4(c + a − b) 4(a − b − c)
p p +1≥ p p − 1.
A+ C B+ C
Write the second inequality as
a2 B − b2 A
p p + (a − b)(a + b + c) ≥ 0.
a B+b A
Since
a2 B − b2 A = (a − b)(a + b + c)(a2 + b2 − 6a b + bc + ca) + 8q(a2 − b2 )

≥ (a − b)(a + b + c)(a2 + b2 − 6a b) ≥ −4a b(a − b)(a + b + c),

−4a b
p p + 1 ≥ 0.
a B+b A
p p p p
Indeed, since A > 8q ≥ 2 a b and B ≥ 8q ≥ 2 a b, we have
p p
p p p p p
a B + b A − 4a b > 2(a + b) a b − 4a b = 2 a b(a + b − 2 a b) ≥ 0.
L
.M
The third inequality holds if
2(a − b − c)
1≥ p p .
B+ C
D
p p
Clearly, it suffices to show that B ≥ a and C ≥ a. We have
A
PI
B − a2 = 8q − 2a(3b − c) + (3b − c)2 ≥ 8a b − 2a(3b − c) = 2a(b + c) ≥ 0
M
and
LY
C − a2 = 8q − 2a(3c − b) + (3c − b)2 ≥ 8a b − 2a(3c − b) = 2a(5b − 3c) ≥ 0.
The equality holds for a = b = c = 1, and also for a = 3 and b = c = 0 (or any cyclic
O
permutation).
.M
9
P 2.60. Let a, b, c be nonnegative real numbers such that a + b + c = 3. If k ≥
W
, then
8
W
p p p p
a2 + bc + k + b2 + ca + k + c 2 + a b + k ≥ 3 2 + k.
W
Solution. We will show that

XÆ XÆ Æ
8(a2 + bc + k) ≥ (3a + b + c)2 + 8k − 9 ≥ 6 2(k + 2).
The right inequality is equivalent to

XÆ Æ
(2a + 3)2 + 8k − 9 ≥ 6 2(k + 2),
and follows immediately from Jensen’s inequality applied to the convex function f :
[0, ∞) → R defined by Æ
f (x) = (2x + 3)2 + 8k − 9.
A1 = 8(a2 + bc + k), B1 = 8(b2 + ca + k), C1 = 8(c 2 + a b + k),
A2 = (3a + b + c)2 + 8k − 9, B2 = (3b + c + a)2 + 8k − 9, C2 = (3c + a + b)2 + 8k − 9,

we can write the left inequality as follows:
A1 − A2 B1 − B2 C1 − C2
p +p p +p p ≥ 0,
A1 + A2 B1 + B2 C1 + C2
p
2bc − ca − a b 2ca − a b − bc 2a b − bc − ca
+ p + p ≥ 0,
L
A1 + A2 B1 + B2 C1 + C2
p p p p
.M
X b(c − a) c(b − a)

p +p ≥ 0,
A1 + A2 A1 + A2
p p
D
X c(a − b) c(b − a)
p +p
A
p ≥ 0,
B1 + B2 A1 + A2
p
X p p p p PI
c(a − b)( C1 + C2 )[( A1 − B1 ) + ( A2 − B2 )] ≥ 0,
p p
M
2(a + b − c) 2a + 2b + c
X
2
c(a − b) ( C1 + C2 ) p p +p
p p
≥ 0.
LY
A1 + B1 A2 + B2
p
Without loss of generality, assume that a ≥ b ≥ c. Clearly, the desired inequality is true
O
for b + c ≥ a. Consider further the case b + c < a. Since a + b − c > 0, it suffices to show
that
.M
2(b + c − a) 2b + 2c + a

2
a(b − c) ( A1 + A2 ) p p +p +
p p
B1 + C1 B2 + C2
p
W
2(c + a − b) 2c + 2a + b

2
+b(a − c) ( B1 + B2 ) p p +p
p p
≥ 0.
W
C1 + A1 C2 + AC2
p
W
Since
b2 (a − c)2 ≥ a2 (b − c)2 ,
2(b + c − a) 2b + 2c + a

b( A1 + A2 ) p p +p +
p p
B1 + C1 B2 + C2
p
2(c + a − b) 2c + 2a + b

+a( B1 + B2 ) p p +p
p p
≥ 0.
C1 + A1 C2 + A2
p
From
a2 B1 − b2 A1 = 8c(a3 − b3 ) + 8k(a2 − b2 ) ≥ 0
and
a2 B2 − b2 A2 = (a − b)(a + b + c)(a2 + b2 + 6a b + bc + ca) + (8k − 9)(a2 − b2 ) ≥ 0,

p p p p
we get a B1 ≥ b A1 and a B2 ≥ b A2 , hence
a( B1 + B2 ) ≥ b( A1 + A2 ).
p p p p
2(b + c − a) 2b + 2c + a 2(c + a − b) 2c + 2a + b
p +p p +p p +p p ≥ 0.
B1 + C1 B2 + C2 C1 + A1 C2 + A2
p
L
.M
This is true if
2b −2b
p +p p ≥0
B1 + C1 C1 + A1
p
D
and
A
−2a 2a 2a
p +p p +p p ≥ 0.
p
B1 + C1 C1 + A1 C2 + A2 PI
The first inequality is true because A1 − B1 = 8(a − b)(a + b − c) ≥ 0. The second
M
inequality can be written as
LY
1 1 1
+p ≥p .
C1 + C2 + B1 +
p p p p
O
A1 A2 C1
Since
.M
1 1 4
+p ≥p ,
C1 + C2 + C1 + A1 + C2 +
p p p p p p
A1 A2 A2
W

W
B1 + 3 C1 ≥ A1 + A2 + C2 .
p p p p p
4
W
Taking into account of

C1 − C2 = 4(2a b − bc − ca) ≥ 0,
C1 − B1 = 8(b − c)(a − b − c) ≥ 0,
A2 − A1 = 4(a b − 2bc + ca) ≥ 0,
we have
B1 + 3 B1 + 2
p p p p p p p p p
4 C1 − A1 − A2 − C2 ≥ 4 C1 − A1 − A2
B1 + 2 B1 −
p p p p
≥4 A2 − A2
= 2(3 B1 − A2 ).
p p
In addition,
9B1 − A2 = 64k − 8a2 + 72b2 − 4a b + 68ac

≥ 72 − 8a2 + 72b2 − 4a b + 68ac
= 8(a + b + c)2 − 8a2 + 72b2 − 4a b + 68ac
= 4(20b2 + 2c 2 + 3a b + 4bc + 21ac) ≥ 0.
Thus, the proof is completed. The equality holds for a = b = c = 1. If k = 9/8, then the
equality holds also for a = 3 and b = c = 0 (or any cyclic permutation).
L
.M
D
p p p p
a3 + 2bc + b3 + 2ca + c 3 + 2a b ≥ 3 3.
A
(Nguyen Van Quy, 2013)
Solution. Since
PI
M
(a3 + 2bc)(a + 2bc) ≥ (a2 + 2bc)2 ,
LY

X a2 + 2bc p
p ≥ 3 3.
O
a + 2bc
By Hölder’s inequality, we have
.M
2 X
a + 2bc
X 2
X 3
p (a2 + 2bc)(a + 2bc) ≥ (a2 + 2bc) = (a + b + c)6 .
a + 2bc
W

W
X
(a + b + c)4 ≥ 3 (a2 + 2bc)(a + 2bc).
W
X
(a + b + c)4 ≥ (a2 + 2bc)(a2 + 6bc + ca + a b).
Indeed,
X X
(a + b + c)4 − (a2 + 2bc)(a2 + 6bc + ca + a b) = 3 a b(a − b)2 ≥ 0.
permutation).

p p p p
a2 + bc b2 + ca c2 + a b 3 2
+ + ≥ .
b+c c+a a+b 2
Solution. According to the well-known inequality
(x + y + z)2 ≥ 3(x y + yz + z x), x, y, z ≥ 0,
L
p
X (b2 + ca)(c 2 + a b 3
≥ .
.M
(c + a)(a + b) 2
Replacing a, b, c by a2 , b2 , c 2 , respectively, the inequality becomes
D
X Æ
A
2 (b2 + c 2 ) (b4 + c 2 a2 )(c 4 + a2 b2 ) ≥ 3(a2 + b2 )(b2 + c 2 )(c 2 + a2 ).
Multiplying the Cauchy-Schwarz inequalities

PI
M
Æ
(b2 + c 2 )(b4 + c 2 a2 ) ≥ b3 + ac 2 ,
LY
Æ
(c 2 + b2 )(c 4 + a2 b2 ) ≥ c 3 + a b2 ,
O
we get Æ
(b2 + c 2 ) (b4 + c 2 a2 )(c 4 + a2 b2 ) ≥ (b3 + ac 2 )(c 3 + a b2 )
.M
= b3 c 3 + a(b5 + c 5 ) + a2 b2 c 2 .
W

W
X X
2 b3 c 3 + 2 a(b5 + c 5 ) + 6a2 b2 c 2 ≥ 3(a2 + b2 )(b2 + c 2 )(c 2 + a2 ).
W

X X X
2 b3 c 3 + 2 bc(b4 + c 4 ) ≥ 3 b2 c 2 (b2 + c 2 ),
X
bc[2b2 c 2 + 2(b4 + c 4 ) − 3bc(b2 + c 2 )] ≥ 0,
X
bc(b − c)2 (2b2 + bc + 2c 2 ) ≥ 0.
P 2.63. If a, b, c are nonnegative real numbers, no two of which are zero,then

p p p
bc + 4a(b + c) ca + 4b(c + a) a b + 4c(a + b) 9
+ + ≥ .
b+c c+a a+b 2
A = bc + 4a(b + c), B = ca + 4b(c + a), C = a b + 4c(a + b).

p
A BC 81
L
X X
+2 ≥ .
(b + c)2 (c + a)(a + b) 4
.M
Further, we use the following identity due to Sung-Yoon Kim:
D
(b + c)2 BC − 4[a(b2 + c 2 ) + 2bc(b + c) + 3a bc]2 = a bc(b − c)2 (a + 4b + 4c),
A
which yields
p
BC ≥
b+c
PI
2a(b2 + c 2 ) + 4bc(b + c) + 6a bc
,
M
p
4 a(b2 + c 2 ) + 8 bc(b + c) + 36a bc
P P
X BC
2 ≥ ,
LY
(c + a)(a + b) (a + b)(b + c)c + a)

p
12 bc(b + c) + 36a bc
P
X BC
O
2 ≥ .
(c + a)(a + b) (a + b)(b + c)c + a)
On the other hand, according to the known inequality Iran-1996,
.M
X a b + bc + ca 9
≥ ,
(b + c) 2 4
W
(see Remark from the proof of P 1.71), we have

W
X A X a b + bc + ca X a 9 X a
= + 3 ≥ + 3 .
(b + c)2 (b + c)2 b+c b+c
W
4
12 bc(b + c) + 36a bc
P
X a
3 + ≥ 18.
b+c (a + b)(b + c)c + a)
This is equivalent to Schur’s inequality of degree three
X X
a3 + 3a bc ≥ bc(b + c).
tion).

p p p
a a2 + 3bc b b2 + 3ca c c 2 + 3a b
+ + ≥ a + b + c.
b+c c+a a+b
(Cezar Lupu, 2006)
Solution. Using the AM-GM inequality, we have

p
a a2 + 3bc 2a(a2 + 3bc) 2a(a2 + 3bc) 2a3 + 6a bc
= p ≥ = ,
b+c 2 (b + c)2 (a2 + 3bc) (b + c)2 + (a2 + 3bc) S + 5bc
L
where S = a2 + b2 + c 2 . Thus, it suffices to show that
.M
X 2a3 + 6a bc
≥ a + b + c.
S + 5bc
D
A
X 2a2 + 6bc
a
S + 5bc
− 1 ≥ 0, PI
M
or, equivalently,
AX + BY + X Z ≥ 0,
LY
where
1 1 1
O
A= , B= , C= ,
S + 5bc S + 5ca S + 5a b
.M
X = a3 + a bc − a(b2 + c 2 ), Y = b3 + a bc − b(c 2 + a2 ), Z = c 3 + a bc − c(a2 + b2 ).

Without loss of generality, assume that a ≥ b ≥ c. We have
W
A ≥ B ≥ C,
W
X = a(a2 − b2 ) + ac(b − c) ≥ 0, Z = c(c 2 − b2 ) + ac(b − a) ≤ 0

W
and, according to Schur’s inequality of third degree,

X X
X +Y +Z = a3 + 3a bc − a(b2 + c 2 ) ≥ 0.
Therefore,
AX + BY + C Z ≥ BX + BY + BZ = B(X + Y + Z) ≥ 0.
tion).

v v v
t 2a(b + c) t 2b(c + a) t 2c(a + b)
+ + ≥ 2.
(2b + c)(b + 2c) (2c + a)(c + 2a) (2a + b)(a + 2b)
p p p
Solution. Making the substitution x = a, y = b, z = c, the inequality becomes
v
X t 2( y 2 + z 2 )
x ≥ 2.
(2 y 2 + z 2 )( y 2 + 2z 2 )
L
We claim that v
2( y 2 + z 2 ) y +z
.M
t
≥ 2 .
(2 y 2 + z )( y + 2z )
2 2 2 y + yz + z 2
Indeed, be squaring and direct calculation, this inequality reduces to y 2 z 2 ( y − z)2 ≥ 0.
D
A
X x( y + z) PI
≥ 2,
y2 + yz + z 2
M
which is just the inequality in P 1.68. The equality holds for a = b = c, and also for
a = 0 and b = c (or any cyclic permutation).
LY
O
.M

v s v v s v
t bc ca t ab t bc ca t ab
+ + ≤ 1 ≤ + + .
W
3a2 + 6 3b2 + 6 3c 2 + 6 6a2 + 3 6b2 + 3 6c 2 + 3

W

W

v 2
a b + bc + ca
X X
Xt bc bc
≤ ,
3a2 + 6 3a2 + 6 a b + bc + ca
hence v 2
Xt bc X 1
≤ .
3a + 6
2 a2 +2
Therefore, to prove the original left inequality, it suffices to show that
X 1
≤ 1.
a2 +2

a2
X
≥ 1.
a2 + 2
Indeed, by the Cauchy-Schwarz inequality, we get
X a2 (a + b + c)2 (a + b + c)2
≥ = = 1.
a2 + 2 (a2 + 2) a2 + 6
P P
The equality occurs for a = b = c = 1. By Hölder’s inequality, we have

v 2
X t bc X X 3
2 2 2
b c (6a + 3) ≥ bc .
6a2 + 3
L
.M
To prove the original right inequality, it suffices to show that
X
(a b + bc + ca)3 ≥ b2 c 2 (6a2 + a b + bc + ca),
D
A
X
(a b + bc + ca)[(a b + bc + ca)2 − PIb2 c 2 ] ≥ 18a2 b2 c 2 ,
2a bc(a b + bc + ca)(a + b + c) ≥ 18a2 b2 c 2 ,

M
X
2a bc a(b − c)2 ≥ 0.
LY
The equality occurs for a = b = c = 1, and for a = 0 and bc = 3 (or any cyclic
permutation).
O
.M
P 2.67. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. If k > 1, than

W
a k (b + c) + b k (c + a) + c k (a + b) ≥ 6.
W
Solution. Let
W
E = a k (b + c) + b k (c + a) + c k (a + b).
We consider two cases.
Case 1: k ≥ 2. Applying Jensen’s inequality to the convex function f (x) = x k−1 , x ≥ 0,
we get
E = (a b + ac)a k−1 + (bc + ba)b k−1 + (ca + c b)c k−1
(a b + ac)a + (bc + ba)b + (ca + c b)c k−1

≥ 2(a b + bc + ca)
2(a b + bc + ca)
k−1
a (b + c) + b2 (c + a) + c 2 (a + b)
2
=6 .
6
a2 (b + c) + b2 (c + a) + c 2 (a + b) ≥ 6.
(a b + bc + ca)(a + b + c) − 3a bc ≥ 6,
a + b + c ≥ 2 + a bc.
It is true since Æ
a+b+c ≥ 3(a b + bc + ca) = 3
L
and
.M
3
a+b+c

a bc ≤ = 1.
3
D
A
Case 2: 1 < k ≤ 2. We have
PI
E = a k−1 (3 − bc) + b k−1 (3 − ca) + c k−1 (3 − a b)
M
= 3(a k−1 + b k−1 + c k−1 ) − a k−1 b k−1 c k−1 (a b)2−k + (bc)2−k + (ca)2−k .

LY
Since 0 ≤ 2 − k < 1, f (x) = x 2−k is concave for x ≥ 0. Thus, by Jensen’s inequality, we

have
O
a b + bc + ca 2−k

(a b)2−k + (bc)2−k + (ca)2−k ≤ 3 = 3,
3
.M
and hence
E ≥ 3(a k−1 + b k−1 + c k−1 ) − 3a k−1 b k−1 c k−1 .
W
Consequently, it suffices to show that

W
a k−1 + b k−1 + c k−1 ≥ a k−1 b k−1 c k−1 + 2.

W
Due to symmetry, we may assume that a ≥ b ≥ c. In addition, write the inequality as
a k−1 + b k−1 − 2 ≥ (a k−1 b k−1 − 1)c k−1 ,

k−1
3 − ab

k−1
a +b k−1
− 2 ≥ (a k−1 k−1
b − 1) .
a+b
Let p p
x= a b, 1 ≤ x ≤ 3.
a + b ≥ 2x, a k−1 + b k−1 ≥ 2x k−1 .


k−1
3 − x2

k−1 2k−2
2(x − 1) ≥ (x − 1) .
2x
Since x ≥ 1, this is true if

k−1
3 − x2

2 ≥ (x k−1
+ 1) ,
2x
L
k−1 k−1
3 − x2 3 − x2

.M
2≥ + .
2 2x
D
Since
3 − x2 3 − x2
A
1≥ ≥ ,
2 2x
PI
the conclusion follows. Thus, the proof is completed. The equality holds for a = b =
c = 1.
M
LY
O
P 2.68. Let a, b, c be nonnegative real numbers such that a + b + c = 2. If 2 ≤ k ≤ 3, than

.M
a k (b + c) + b k (c + a) + c k (a + b) ≤ 2.
W
Solution. Denote by Ek (a, b, c) the left hand side of the inequality, assume that a ≤ b ≤
W
c and show that

Ek (a, b, c) ≤ Ek (0, a + b, c) ≤ 2.
W
The left inequality is equivalent to
a b k−1
(a + b k−1 ) ≤ (a + b)k − a k − b k .
c
Clearly, it suffices to consider c = b, when the inequality becomes
2a k + b k−1 (a + b) ≤ (a + b)k .
Since 2a k ≤ a k−1 (a + b), it remains to show that
a k−1 + b k−1 ≤ (a + b)k−1 ,

which is true since

k−1
a k−1 + b k−1 a k−1

b a b
= + ≤ + = 1.
(a + b) k−1 a+b a+b a+b a+b
The right inequality, Ek (0, a + b, c) ≤ 2, is equivalent to
cd(c k−1 + d k−1 ) ≤ 2,
where d = a + b, and hence c + d = 2. By the Power-Mean inequality (or Jensen’s

k−1
inequality applied to the concave function t 2 ), we have
L
1/(k−1) 1/2
c k−1 + d k−1 c2 + d 2

.M
≤ ,
2 2
D
(k−1)/2
c2 + d 2

k−1
c +d k−1
≤2 .
A
2
Thus, it suffices to show that PI
M
(k−1)/2
c2 + d 2

cd ≤ 1,
LY
O
cd(2 − cd)(k−1)/2 ≤ 1.
.M
Since 2 − cd ≥ 1, we have
cd(2 − cd)(k−1)/2 ≤ cd(2 − cd) = 1 − (1 − cd)2 ≤ 1.

W

W
W
P 2.69. Let a, b, c be nonnegative real numbers, no two of which are zero. If m > n ≥ 0,
than
(b + c − 2a) + (c + a − 2b) + (a + b − 2c) ≥ 0.
AX + BY + C Z ≥ 0,
where
A= , B = , C = ,
X = b + c − 2a, Y = c + a − 2b, Z = a + b − 2c, X + Y + Z = 0.
Without loss of generality, assume that a ≤ b ≤ c, which involves X ≥ Y ≥ Z and X ≥ 0.
Since
2(AX + BY + C Z) = (2A − B − C)X + (B + C)X + 2(BY + C Z)

= (2A − B − C)X − (B + C)(Y + Z) + 2(BY + C Z)
= (2A − B − C)X + (B − C)(Y − Z),
L
it suffices to show that B ≥ C and 2A − B − C ≥ 0. The inequality B ≥ C can be written
.M
as
b n c n (c m−n − b m−n ) + a n (c m − b m ) − a m (c n − b n ) ≥ 0,
D
b n c n (c m−n − b m−n ) + a n [c m − b m − a m−n (c n − b n )] ≥ 0.
A
This is true since c m−n ≥ b m−n and
PI
c m − b m − a m−n (c n − b n ) ≥ c m − b m − b m−n (c n − b n ) = c n (c m−n − b m−n ) ≥ 0.
M
The inequality 2A − B − C ≥ 0 follows from
LY
2A ≥ b m−n + c m−n , b m−n ≥ C, c m−n ≥ B.

O
Indeed, we have
.M
(b n − c n )(b m−n − c m−n )

2A − b m−n − c m−n = ≥ 0,
bn + c n
W
a n (b m−n − a m−n )
b m−n − C = ≥ 0,
W
an + bn
a n (c m−n − a m−n )
c m−n − B =
W
≥ 0.
c n + an
tion).

p p p
a2 − a + 1 + b2 − b + 1 + c 2 − c + 1 ≥ a + b + c.

First Solution. Among a − 1, b − 1 and c − 1 there are two with the same sign. Let
(b − 1)(c − 1) ≥ 0, that is,
1
t ≤ , t = b + c − 1.
a
By Minkowsky’s inequality, we have
v v
2
1 2 3
u u
1 3 t
p p
b − b+1+ c −c+1= + + +
t
2 2 b− c−
2 4 2 4
p
≥ t 2 + 3.
L
.M
p p
a2 − a + 1 + t 2 + 3 ≥ a + b + c,
D
p
a2 − a + 1 + f (t) ≥ a + 1,
A
where
f (t) =
p PI
t 2 + 3 − t.
M
Clearly, f (t) is decreasing for t ≤ 0. Since
LY
3
f (t) = p ,
t2 + 3 + t
O
1

f (t) is also decreasing for t ≥ 0. Then, f (t) ≥ f , and it suffices to show that
a
.M
1
p
a2 − a + 1 + f ≥ a + 1,
a
W
W
v
p t1 1
a2 − a + 1 + + 3 ≥ a + + 1.
W
a2 a
v
1 2
t
2 (a − a + 1) 2 + 3 ≥ 3a + − 1.
2
a a
Indeed, by the Cauchy-Schwarz inequality, we have
v v
1 1
t t
2 (a − a + 1) 2 + 3 = [(2 − a)2 + 3a2 ] 2 + 3
2
a a
2−a 2
≥ + 3a = 3a + − 1.
a a

Second Solution. If the inequality
1 3
p
x2 − x +1− x ≥ −1
2 x + x +1
2
holds for all x > 0, then it suffices to prove that
1 1 1
+ + ≥ 1,
a2 + a + 1 b2 + b + 1 c 2 + c + 1
which is just the known inequality in P 1.44. Indeed, the above inequality is equivalent
L
to
.M
1− x (1 − x)(2 + x)
p ≥ ,
x − x +1+ x
2 2(x 2 + x + 1)
D
p
(x − 1)[(x + 2) x 2 − x + 1 − x 2 − 2] ≥ 0,
A
3x 2 (x − 1)2
p ≥ 0.
(x + 2) x 2 − x + 1 + x 2 + 2 PI
M
LY

O
p p p
16a2 + 9 + 16b2 + 9 + 16b2 + 9 ≥ 4(a + b + c) + 3.
.M
(MEMO, 2012)
W
First Solution (by Vo Quoc Ba Can). Since
9
W
p
16a2 + 9 − 4a = p ,
16a2 + 9 + 4a
W

X 1 1
p ≥ .
16a2 + 9 + 4a 3

p 16a2 + 9
2 16a2 + 9 ≤ + 2a + 3,
2a + 3
p 16a2 + 9 18(2a2 + 2a + 1)
2( 16a2 + 9 + 4a) ≤ + 10a + 3 = .
2a + 3 2a + 3

X 2a + 3
≥ 3.
2a2 + 2a + 1
If the inequality
2a + 3 3
≥ 8/5
2a2 + 2a + 1 a + a4/5 + 1
holds for all a > 0, then it suffices to show that
X 1
≥ 1,
a + a4/5 + 1
8/5
L
which follows immediately from the inequality in P 1.44. Therefore, using the substitu-
.M
tion x = a1/5 , x > 0, we need to show that
2x 5 + 3 3
D
≥ 8 ,
2x + 2x + 1
10 5 x + x4 + 1
A
PI
2x 4 (x 5 − 3x 2 + x + 1) + x 4 − 4x + 3 ≥ 0.
M
This is true since, by the AM-GM inequality, we have
LY
3
p
x 5 + x + 1 ≥ 3 x 5 · x · 1 = 3x 2
O
and
4
p
x4 + 3 = x4 + 1 + 1 + 1 ≥ 4 x 4 · 1 · 1 · 1 = 4x.
.M

W
Second Solution. Making the substitution

p p p
W
x = 16a2 + 9 − 4a, y = 16b2 + 9 − 4b, z = 16c 2 + 9 − 4c, x, y, z > 0,

W
which involves
9 − x2 9 − y2 9 − z2
a= , b= , c= ,
8x 8y 8z
(9 − x 2 )(9 − y 2 )(9 − z 2 ) = 512x yz
yields
x + y + z ≥ 3.
Use the contradiction method. Assume that x + y + z < 3 and show that
(9 − x 2 )(9 − y 2 )(9 − z 2 ) > 512x yz.

According to the AM-GM inequality, we get

p p p
3 + x = 1 + 1 + 1 + x ≥ 4 x, 3 + y ≥ 4 4 y, 3 + z ≥ 4 z,
4 4
hence
p
(3 + x)(3 + y)(3 + z) ≥ 64 4 x yz.
(3 − x)(3 − y)(3 − z) > 8(x yz)3/4 .

x + y + z 3
L
1> ≥ x yz,
.M
3
and hence
D
(3 − x)(3 − y)(3 − z) = 9(3 − x − y − z) + 3(x y + yz + z x) − x yz
A
> 3(x y + yz + z x) − x yz ≥ 9(x yz)2/3 − x yz
PI
> 8(x yz)2/3 > 8(x yz)3/4 .
M
LY

O
p p p
25a2 + 144 + 25b2 + 144 + 25c 2 + 144 ≤ 5(a + b + c) + 24.
.M

W

144
W
p
25a2 + 144 − 5a = p ,
25a2 + 144 + 5a
W

X 1 1
p ≤ .
25a2 + 144 + 5a 6
If the inequality
1 1
p ≤ p
25a2 + 144 + 5a 6 5a18/13 + 4
holds for all a > 0, then it suffices to show that
X 1
p ≤ 1,
5a18/13 + 4
which follows immediately from P 2.31. Therefore, using the substitution x = a1/13 ,
x > 0, we only need to show that
p p
25x 26 + 144 + 5x 13 ≥ 6 5x 18 + 4.

p
10x 13 ( 25x 26 + 144 + 5x 13 − 18x 5 ) ≥ 0.
This is true if
25x 26 + 144 ≥ (18x 5 − 5x 13 )2 ,
L
.M
5x 18 + 4 ≥ 9x 10 .
D
5x 18 + 4 = x 18 + x 18 + x 18 + x 18 + x 18 + 1 + 1 + 1 + 1
A
9
p
≥ 9 x 18 · x 18 · x 18 · x 18 · x 18 · 1 · 1 · 1 · 1 = 9x 10 .
PI
M
Second Solution. Making the substitution
LY
p p p
8x = 25a2 + 144 − 5a, 8 y = 25b2 + 144 − 5b, 8z = 25c 2 + 144 − 5c,
O
which involves
.M
9 − 4x 2 9 − 4 y2 9 − 4z 2 3

a= , b= , c= , x, y, z ∈ 0, ,
5x 5y 5z 2
W

W
(9 − 4x 2 )(9 − 4 y 2 )(9 − 4z 2 ) = 125x yz

W
involves
x + y + z ≤ 3.
Use the contradiction method. Assume that x + y + z > 3 and show that
(9 − 4x 2 )(9 − 4 y 2 )(9 − 4z 2 ) < 125x yz.
Since
12x 2 3( y + z − x)( y + z + 3x)
9 − 4x 2 < 3(x + y + z) − = ,
x + y +z x + y +z
27AB ≤ 125x yz(x + y + z)3 ,
where
A = ( y + z − x)(z + x − y)(x + y − z), B = ( y + z + 3x)(z + x + 3 y)(x + y + 3z).
Consider the nontrivial case A ≥ 0. By the AM-GM inequality, we have
125
B≤ (x + y + z)3 .
27
A ≤ x yz,
which is a well known inequality (equivalent to Schur’s inequality of degree three).
L
D
.M
P 2.73. If a, b are positive real numbers such that a b + bc + ca = 3, then
A
p p p
(a) a2 + 3 + b2 + 3 + b2 + 3 ≥ a + b + c + 3;
p p p p
PI
(b) a+b+ b+c+ c+a≥ 4(a + b + c) + 6.
M
(Lee Sang Hoon, 2007)
LY
Solution. (a) First Solution (by Pham Thanh Hung). By squaring, the inequality be-
O
comes XÆ
(b2 + 3)(c 2 + 3) ≥ 3(1 + a + b + c).
.M
Since
1
(b2 + 3)(c 2 + 3) = (b + c)(b + a)(c + a)(c + b) = (b + c)2 (a2 + 3) ≥ (b + c)2 (a + 3)2 ,
W
4
W
we have
1X 1 X
W
XÆ X
(b2 + 3)(c 2 + 3) ≥ (b + c)(a + 3) = 2 bc + 6 a = 3(1 + a + b + c).
2 2
Second Solution. Write the inequality as follows:
Æ Æ Æ
(a + b)(a + c) + (b + c)(b + a) + (c + a)(c + b) ≥ a + b + c + 3,
Æ X p p 2
2 a + b + c − 3(a b + bc + ca) ≥ a+b− a+c ,
1 X X (b − c)2
p (b − c)2 ≥ p p 2 ,
a+b+c+ 3(a b + bc + ca) a+b+ a+c
X Sa (b − c)2
p p 2 ≥ 0,
a+b+ a+c
where p p 2 Æ
Sa = a+b+ a+c −a−b−c− 3(a b + bc + ca).
Æ Æ
Sa = 3(a + b + c) + 2 (a + b)(a + c) − 3(a b + bc + ca)
Æ Æ
>2 a2 + (a b + bc + ca) − 3(a b + bc + ca) > 0.
L
Third Solution. Use the substitution
.M
p p p
x = a2 + 3 − a, y = b2 + 3 − b, z = c 2 + 3 − c, x, y, z > 0.
D
A
x + y + z ≥ 3.
We have XX Æ Æ
PI
yz =
[ (b + a)(b + c) − b][ (c + a)(c + b) − c]
M
X Æ X Æ X Æ X
= (b + c) (a + b)(a + c) − b (c + a)(c + b) − c (b + a)(b + c) +
LY
bc
X
= bc = 3.
O
Thus, we get
.M
Æ
x + y +z ≥ 3(x y + yz + z x) = 3.
(b) By squaring, we get the inequality in (a). Otherwise, using the substitution
W
p p p
x= b + c, y= c + a, z = a + b,
W

W
Ç Æ
x + y +z ≥ 2(x 2 + y 2 + z 2 ) + 3(2x 2 y 2 + 2 y 2 z 2 + 2z 2 x 2 − x 4 − y 4 − z 4 ).
By squaring two times, we get

Æ
2(x y + yz + z x) − x 2 − y 2 − z 2 ≥ 3(2x 2 y 2 + 2 y 2 z 2 + 2z 2 x 2 − x 4 − y 4 − z 4 ),
X
(x − y)2 (x + y − z)2 ≥ 0.

Æ Æ Æ
(5a2 + 3)(5b2 + 3) + (5b2 + 3)(5c 2 + 3) + (5c 2 + 3)(5a2 + 3) ≥ 24.
(Nguyen Van Quy, 2012)
Solution. Assume that a ≥ b ≥ c, which involves 1 ≤ a ≤ 3 and b + c ≤ 2. Denote
A = 5a2 + 3, B = 5b2 + 3, C = 5c 2 + 3,

p p p p
L
A ( B + C) + BC ≥ 24,
.M
p q p p
A · A(B + C + 2 BC ) ≥ 24 − BC.
D
This is true if p p
A
A(B + C + 2 BC) ≥ (24 − BC)2 ,
PI
p
M
A(A + B + C + 48) ≥ (A + 24 − BC)2 .
LY
Applying Lemma below for k = 5/3 and m = 4/15 yields

p
O
5 BC ≥ 4(b − c)2 + 25bc + 15.

.M
25A(A + B + C + 48) ≥ [5A + 120 − 4(b − c)2 − 25bc − 15]2 ,

W
W
25(5a2 + 3)[5(a2 + b2 + c 2 ) + 57] ≥ [25a2 + 120 − 4(b − c)2 − 25bc]2 .

W
Since
5(a2 + b2 + c 2 ) + 57 = 5a2 + 5(b + c)2 − 10bc + 57 = 2(5a2 − 15a + 51 − 5bc)
and
25a2 + 120 − 4(b − c)2 − 25bc = 25a2 + 120 − 4(b + c)2 − 9bc = 3(7a2 + 8a + 28 − 3bc),
50(5a2 + 3)(5a2 − 15a + 51 − 5bc) ≥ 9(7a2 + 8a + 28 − 3bc)2 .

From bc ≤ (b + c)2 /4 and (a − b)(a − c) ≥ 0, we get
bc ≤ (3 − a)2 /4, bc ≥ a(b + c) − a2 = 3a − 2a2 .
Consider a fixed, a ≥ 1, and denote x = bc. So, we only need to prove that f (x) ≥ 0
for
a2 − 6a + 9
3a − 2a2 ≤ x ≤ ,
4
where
f (x) = 50(5a2 + 3)(5a2 − 15a + 51 − 5x) − 9(7a2 + 8a + 28 − 3x)2 .
L
a − 6a + 9
2
Since f is concave, it suffices to show that f (3a − 2a ) ≥ 0 and f
2
.M
≥ 0.
4
Indeed,
D
f (3a − 2a2 ) = 3(743a4 − 2422a3 + 2813a2 − 1332a + 198)
A
= 3(a − 1)2 [(a − 1)(743a − 193) + 5] ≥ 0,
PI
a2 − 6a + 9

M
375
f = (25a4 − 140a3 + 286a2 − 252a + 81)
4 16
LY
375
= (a − 1)2 (5a − 9)2 ≥ 0.
16
O
Thus, the proof is completed. The equality holds for a = b = c = 1, and also for a = 9/5
.M

k
Lemma. Let b, c ≥ 0 such that b + c ≤ 2. If k > 0 and 0 ≤ m ≤ , then
2k + 2
W
Æ
(k b2 + 1)(kc 2 + 1) ≥ m(b − c)2 + k bc + 1.
W

W
(b − c)2 [k − 2m − 2kmbc − m2 (b − c)2 ] ≥ 0.
This is true since
k − 2m − 2kmbc − m2 (b − c)2 = k − 2m − 2m(k − 2m)bc − m2 (b + c)2

m(k − 2m)
≥ k − 2m − (b + c)2 − m2 (b + c)2
2
km
= k − 2m − (b + c)2 ≥ k − 2m − 2km ≥ 0.
2

v
p p p t 4(a2 + b2 + c 2 ) + 42
a2 + 1 + b2 + 1 + c2 + 1 ≥ .
3
Solution. Assume that a ≥ b ≥ c, which involves a ≥ 1 and b + c ≤ 2. By squaring, the

inequality becomes
p p p p a2 + b2 + c 2 + 33
A ( B + C) + BC ≥ ,
L
6
.M
p p a2 + b2 + c 2 + 33
q
A(B + C + 2 BC ) + BC ≥ ,
6
D
where
A
A = a2 + 1, B = b2 + 1, C = c 2 + 1.
PI
Applying Lemma from the preceding problem P 2.74 for k = 1 and m = 1/4 gives
M
p 1
BC ≥ (b − c)2 + bc + 1.
4
LY

O
v
t 1 1 a2 + b2 + c 2 + 33
.M
A[B + C + (b − c)2 + 2bc + 2] + (b − c)2 + bc + 1 ≥ ,

2 4 6
W
Æ
6 2(a2 + 1)[3(b + c)2 + 8 − 4bc] ≥ 2a2 − (b + c)2 + 54 − 4bc,
W
Æ
W
6 2(a2 + 1)(3a2 − 18a + 35 − 4bc) ≥ a2 + 6a + 45 − 4bc.
bc ≤ (3 − a)2 /4, bc ≥ a(b + c) − a2 = 3a − 2a2 .
for
a2 − 6a + 9
3a − 2a2 ≤ x ≤ ,
4
where
f (x) = 72(a2 + 1)(3a2 − 18a + 35 − 4x) − (a2 + 6a + 45 − 4x)2 .
a2 − 6a + 9

2
≥ 0.
4
Indeed,
f (3a − 2a2 ) = 9(79a4 − 228a3 + 274a2 − 180a + 55)

= 9(a − 1)2 (79a2 − 70a + 55 ≥ 0,
a2 − 6a + 9

f = 144(a4 − 6a3 + 13a2 − 12a + 4)
4
= 144(a − 1)2 (a − 2)2 ≥ 0.
L
.M
The equality holds for a = b = c = 1, and also for a = 2 and b = c = 1/2 (or any cyclic
permutation).
D
A
PI
M
p p p p
(a) a2 + 3 + b2 + 3 + c 2 + 3 ≥ 2(a2 + b2 + c 2 ) + 30;
LY
p p p p
(b) 3a2 + 1 + 3b2 + 1 + 3c 2 + 1 ≥ 2(a2 + b2 + c 2 ) + 30.
O
Solution. Assume that a ≥ b ≥ c, which involves a ≥ 1 and b + c ≤ 2.

.M
(a) By squaring, the inequality becomes

W
p p p p a2 + b2 + c 2 + 21
A ( B + C) + BC ≥ ,
2
W
p p a2 + b2 + c 2 + 21
q
A(B + C + 2 BC ) + BC ≥ ,
W
2
where
A = a2 + 3, B = b2 + 3, C = c 2 + 3.
Applying Lemma from problem P 2.74 for k = 1/3 and m = 1/9 gives
p 1
BC ≥ (b − c)2 + bc + 3.
3
v
t 2 1 a2 + b2 + c 2 + 21
A[B + C + (b − c)2 + 2bc + 6] + (b − c)2 + bc + 3 ≥ ,
3 3 2
Æ
2 3(a2 + 3)[5(b + c)2 + 36 − 8bc] ≥ 3a2 + (b + c)2 + 45 − 4bc,
Æ
3(a2 + 3)(5a2 − 30a + 81 − 8bc) ≥ 2a2 − 3a + 27 − 2bc.
bc ≤ (3 − a)2 /4, bc ≥ a(b + c) − a2 = 3a − 2a2 .
L
for
a2 − 6a + 9
.M
3a − 2a2 ≤ x ≤ ,
4
where
D
f (x) = 3(a2 + 3)(5a2 − 30a + 81 − 8x) − (2a2 − 3a + 27 − 2x)2 .
A
a − 6a + 9
2
2
PI 4
≥ 0.
Indeed,
M
f (3a − 2a2 ) = 27a2 (a − 1)2 ≥ 0,
LY
a2 − 6a + 9

27 4
= (a − 8a3 + 22a2 − 24a + 9)
O
f
4 4
27
.M
= (a − 1)2 (a − 3)2 ≥ 0.
4
W
permutation).
W

W
p p p p 27 − a2 − b2 − c 2
A ( B + C) + BC ≥ ,
2
p p 27 − a2 − b2 − c 2
q
A(B + C + 2 BC ) + BC ≥ ,
2
where
A = 3a2 + 1, B = 3b2 + 1, C = 3c 2 + 1.
Applying Lemma from problem P 2.74 for k = 3 and m = 1/3 gives

p 1
BC ≥ (b − c)2 + 3bc + 1.
3

v
t 2 1 27 − a2 − b2 − c 2
A[B + C + (b − c)2 + 6bc + 2] + (b − c)2 + 3bc + 1 ≥ ,
3 3 2
Æ
2 3(3a2 + 1)[11(b + c)2 + 12 − 8bc] ≥ 75 − 3a2 − 5(b + c)2 − 4bc,
Æ
3(3a2 + 1)(11a2 − 66a + 111 − 8bc) ≥ 15 + 15a − 4a2 − 2bc.
L
bc ≤ (3 − a)2 /4, bc ≥ a(b + c) − a2 = 3a − 2a2 .
.M
D
for
a2 − 6a + 9
3a − 2a2 ≤ x ≤
A
,
4
where PI
M
f (x) = 3(3a2 + 1)(11a2 − 66a + 111 − 8x) − (15 + 15a − 4a2 − 2x)2 .
LY
a − 6a + 9
2
2
≥ 0.
4
O
Indeed,
f (3a − 2a2 ) = 27(a − 1)2 (3a − 2)2 ≥ 0,
.M
a2 − 6a + 9

27
f = (9a4 − 48a3 + 94a2 − 80a + 25)
4 4
W
27
= (a − 1)2 (3a − 5)2 ≥ 0.
W
4
The equality holds for a = b = c = 1, and also for a = 5/3 and b = c = 2/3 (or any
W
• Let a, b, c be nonnegative real numbers such that a + b + c = 3. If k > 0, then
v
p p p t 8k(a2 + b2 + c 2 ) + 3(9k2 + 10k + 9)
ka2 + 1 + k b2 + 1 + kc 2 + 1 ≥ ,
3(k + 1)
3k + 1 3k − 1
with equality for a = b = c = 1, and also for a = and b = c = (or any
2k 4k

Æ Æ Æ
(32a2 + 3)(32b2 + 3) + (32b2 + 3)(32c 2 + 3) + (32c 2 + 3)(32a2 + 3) ≤ 105.
Solution. Assume that a ≤ b ≤ c, which involves a ≤ 1 and b + c ≥ 2. Denote
A = 32a2 + 3, B = 32b2 + 3, C = 32c 2 + 3,

p p p p
L
A ( B + C) + BC ≤ 105,
.M
p Æ p p
A · B + C + 2 BC ≤ 105 − BC.
D
By Lemma below, we have
A
p
BC ≤ 5(b + c)2 + 12bc + 3 ≤ 8(b + c)2 + 3 ≤ 8(a + b + c)2 + 3 = 75 < 105.
Therefore, we can write the desired inequality as

PI
M
p p
A(B + C + 2 BC) ≤ (105 − BC)2 ,
LY
O
p
A(A + B + C + 210) ≤ (A + 105 − BC)2 .
.M
According to Lemma below, it suffices to show that
A(A + B + C + 210) ≤ [A + 105 − 5(b2 + c 2 ) − 22bc − 3]2 ,

W
W
[32a2 + 105 − 5(b2 + c 2 ) − 22bc]2 ≥ (32a2 + 3)[32(a2 + b2 + c 2 ) + 219].

W
Since
32(a2 + b2 + c 2 ) + 219 = 32a2 + 32(b + c)2 − 64bc + 219 = 64a2 − 192a + 507 − 64bc
and
32a2 +105−5(b2 +c 2 )−22bc = 32a2 +105−5(b+c)2 −12bc = 3(9a2 +10a+20−4bc),
9(9a2 + 10a + 20 − 4bc)2 ≥ (32a2 + 3)(64a2 − 192a + 507 − 64bc).

From bc ≤ (b + c)2 /4, we get

bc ≤ (3 − a)2 /4.
Consider a fixed, 0 ≤ a ≤ 1, and denote x = bc. So, we only need to prove that f (x) ≥ 0
for
a2 − 6a + 9
x≤ ,
4
where
f (x) = 9(9a2 + 10a + 20 − 4x)2 − (32a2 + 3)(64a2 − 192a + 507 − 64x).
Since
L
f 0 (x) = 72(4x − 9a2 − 10a − 20) + 64(32a2 + 3)
.M
≤ 72[(a2 − 6a + 9) − 9a2 − 10a − 20) + 64(32a2 + 3)
= 8[184a(a − 1) + (44a − 75)] < 0,
D
A
a − 6a + 9
2
f is decreasing, hence f (x) ≥ f . Therefore, it suffices to show that
a − 6a + 9
2 4 PI
f ≥ 0. We have
M
4
LY
a − 6a + 9
2
f =9[9a2 + 10a + 20 − (a2 − 6a + 9)]2
4
O
− (32a2 + 3)[64a2 − 192a + 507 − 16(a2 − 6a + 9)]

.M
=9(8a2 + 16a + 11)2 − (32a2 + 3)(48a2 − 96a + 363)

=192a(a − 1)2 (18 − 5a) ≥ 0.
W
Thus, the proof is completed. The equality holds for a = b = c = 1, and also for a = 0
W
Lemma. If b, c ≥ 0 such that b + c ≥ 2, then

W
Æ
(32b2 + 3)(32c 2 + 3) ≤ 5(b2 + c 2 ) + 22bc + 3.
(5b2 + 5c 2 + 22bc)2 − 322 b2 c 2 ≥ 96(b2 + c 2 ) − 6(5b2 + 5c 2 + 22bc),
5(b − c)2 (5b2 + 5c 2 + 54bc) ≥ 66(b − c)2 .

5(5b2 + 5c 2 + 10bc) ≥ 100,
which is equivalent to the obvious inequality (b + c)2 ≥ 4.

b+c

c + a
+

a b
a − 3 + b − 3 + c − 3 ≥ 2.

Solution. Without loss of generality, assume that a ≥ b ≥ c.

Case 1: a > b + c. We have
b+c a + b c + a b+c b+c

a − 3 + c − 3 + b − 3 ≥ a − 3 = 3 − a > 2.

L
.M
Case 2: a ≤ b + c. We have
b+c a + b c + a b+c c + a

D

− 3 + − 3 + − 3 ≥ − 3 + − 3

a c
b a b
A
b+c c + a 2b b + a (a − b)(2b − a)
PI

= 3− + 3− ≥6− − =2+ ≥ 2.
a b a b ab
M
a
Thus, the proof is completed. The equality holds for = b = c (or any cyclic permuta-
2
LY
tion).
O
.M
P 2.79. If a, b, c are real numbers such that a bc 6= 0, then
b + c c + a a + b

W
+ +
≥ 2.
a b c
W
W
First Solution. Let

|a| = max{|a|, |b|, |c|}.
We have
b + c c + a a + b b + c c + a a + b

+ +
≥ + +

a b c a a a
|(−b − c) + (c + a) + (a + b)|
≥ = 2.
|a|
The equality holds for a = 1, b = −1 and |c| ≤ 1 (or any permutation).
Second Solution. Since the inequality remains unchanged by replacing a, b, c with
−a, −b, −c, it suffices to consider two cases: a, b, c > 0, and a < 0 and b, c > 0.
Case 1: a, b, c > 0. We have

b + c c + a a + b a b b c c a

+ +
= + + + + + ≥ 6.
a b c b a c b a c
Case 2: a < 0 and b, c > 0. Replacing a by −a, we need to show that
b + c |a − c| |a − b|
+ + ≥2
a b c
for all a, b, c > 0. Without loss of generality, assume that b ≥ c.
For b ≥ c ≥ a, we have
L
.M
b + c |a − c| |a − b| b+c
+ + ≥ ≥ 2.
a b c a
D
For b ≥ a ≥ c, we have
A
b + c |a − c| |a − b| b+c a−c (a − b)2 + c(b − a)
+ + −2≥ + −2= PI ≥ 0.
a b c a b ab
M
For a ≥ b ≥ c, we have
b + c |a − c| |a − b| b+c a−c a−b
LY
+ + = + +
a b c a b c
O
a−b 1 1 (a − b)2 (a − b)(a b − c 2 )

a b
= + −2 + +c − = + ≥ 0.
b a c a b ab a bc
.M
Third Solution. Using the substitution

W
b+c c+a a+b

x= , y= , z= ,
W
a b c
W
x + y + z + 2 = x yz, x, y, z ∈ R
involves
|x| + | y| + |z| ≥ 2.
If x yz ≤ 0, then
−x − y − z = 2 − x yz ≥ 2,
hence
|x| + | y| + |z| ≥ |x + y + z| = | − x − y − z| ≥ −x − y − z ≥ 2.
If x yz > 0, then either x, y, z > 0 or only one of x, y, z is positive (for instance, x > 0
and y, z < 0).
Case 1: x, y, z > 0. We need to show that x + y + z ≥ 2. We have
x yz = x + y + z + 2 > 2
and, by the AM-GM inequality, we get

p p
3
x + y + z ≥ 3 3 x yz > 3 2 > 2,
Case 2: x > 0 and y, z < 0. Replacing y, z by − y, −z, we need to prove that
L
x − y − z + 2 = x yz
.M
involves
x + y +z ≥2
D
for all x, y, z > 0. Since
A
PI
x + y + z − 2 = x + y + z − (x yz − x + y + z) = x(2 − yz),
M
we need to show that yz ≤ 2. Indeed, we have
p
LY
x + 2 = y + z + x yz ≥ 2 yz + x yz,
p
O
x(1 − yz) + 2(1 − yz ) ≥ 0,

p p
(1 − yz )[x(1 + yz ) + 2] ≥ 0,
.M
hence
yz ≤ 1 < 2.
W
W
W
2a 2b 2c
x= , y= , z= .
b+c c+a a+b
Prove that
p p p
(a) x + y +z+ xy+ yz + z x ≥ 6;
p p p
x+ y+ z ≥ 8 + x yz.
p
(b)
Solution. (a) First Solution. Since

p p p
p 2 bc(a + b)(c + a) 2 bc (a + bc)
yz = ≥
(a + b)(c + a) (a + b)(c + a)
p
2a(b + c) bc + 2bc(b + c) 4a bc + 2bc(b + c)
= ≥ ,
(a + b)(b + c)(c + a) (a + b)(b + c)(c + a)
we have
12a bc + 2 bc(b + c)
P
Xp
yz ≥
(a + b)(b + c)(c + a)
8a bc
L
= + 2 = x yz + 2.
(a + b)(b + c)(c + a)
.M
D
x + y + z + x yz ≥ 4,
A
a3 + b3 + c 3 + 3a bc ≥
X PI
a b(a + b).
M
LY
tion).
O
X Xp p
4 (x − 1) ≥ ( y − z)2 .
.M
Since
X X (a − b) + (a − c) X a − b X b − a X (a − b)2
W
(x − 1) = = + =
b+c b+c c+a (b + c)(c + a)
W
X (b − c)2
= ,
(a + b)(a + c)
W
and
p p 2(b − c)2 (a + b + c)2
( y − z)2 = p p 2 ,
(a + b)(a + c) b2 + a b + c 2 + ac
X
(b − c)2 Ea ≥ 0,
where  
(a + b + c) 2
Ea = (b + c) 2 − p p 2  .
b + a b + c + ac
2 2
By Minkowski’s inequality, we have

p p 2 p p
b2 + a b + c 2 + ac ≥ (b+c)2 +a( b+ c)2 ≥ (b+c)2 +a(b+c) = (b+c)(a+b+c).
Therefore,
a+b+c

Ea ≥ (b + c) 2 − = b + c − a,
b+c
and hence X X
(b − c)2 Ea ≥ (b − c)2 (b + c − a) ≥ 0.
The right inequality is just Schur’s inequality of third degree.
L
Third Solution. Using the Cauchy-Schwarz inequality, we have
.M
a b c (a + b + c)2 (a + b + c)2
+ + ≥ = .
b+c c+a a+b a(b + c) + b(c + a) + c(a + b) 2(a b + bc + ca)
D
A
Using Hölder’s inequality, we have
s
a
v
t b s
c
2 PI
(a + b + c)3
+ + ≥ 2 .
M
b+c c+a a+b a (b + c) + b2 (c + a) + c 2 (a + b)
LY

O
(a + b + c)2 2(a + b + c)3

+ 2 ≥ 12.
a b + bc + ca a (b + c) + b2 (c + a) + c 2 (a + b)
.M
Due to homogeneity, we may assume that a + b + c = 1. Substituting q = a b + bc + ca,

3q ≤ 1, the inequality becomes
W
1 2
+ ≥ 12.
W
q q − 3a bc
W
From the fourth degree Schur’s inequality
6a bc p ≥ (p2 − q)(4q − p2 ), p = a + b + c,
we obtain
6a bc ≥ (4q − 1)(1 − q).
Therefore,
1 2 1 2
+ − 12 ≥ + (4q−1)(1−q)
− 12
q q − 3a bc q q−
2
1 4 (1 − 3q)(1 − 4q)2
= + 2 − 12 = ≥ 0.
q 4q − 3q + 1 q(4q2 − 3q + 1)

p p p
x + y + z + 2 x y + 2 yz + 2 z x ≥ 8 + x yz.
As we have shown in the first solution of the inequality in (a),

p p p
x y + yz + z x ≥ x yz + 2.
x + y + z + 2(x yz + 2) ≥ 8 + x yz,
L
x + y + z + x yz ≥ 4.
.M
This inequality reduces to Schur’s inequality of third degree. The equality holds for
a = b = c, and also for a = 0 and b = c (or any cyclic permutation).
D
A
PI
M
2a 2b 2c
x= , y= , z= .
LY
b+c c+a a+b

Prove that
O
p p p
1 + 24x + 1 + 24 y + 1 + 24z ≥ 15.
.M
Solution (by Vo Quoc Ba Can). Assume that c = min{a, b, c}, hence z = min{x, y, z},
W
z ≤ 1. By Hölder’s inequality
2
W
s v
a t b 2
a (b + c) + b2 (c + a) ≥ (a + b)3 ,

+
b+c c+a
W
we get
p p 2 2(a + b)3 2(a + b)3

x+ y ≥ ≥
c(a + b)2 + a b(a + b − 2c) c(a + b)2 + 14 (a + b)2 (a + b − 2c)
8(a + b) 8
= = .
a + b + 2c 1+z
Using this result and Minkowski’s inequality, we have
v
p p q p p 2 t 48
1 + 24x + 1 + 24 y ≥ (1 + 1)2 + 24( x + y) ≥ 2 1 + .
1+z

v
t 48 p
2 1+ + 1 + 24z ≥ 15.
1+z

p 1
1 + 24z = 5t, ≤ t ≤ 1,
5
the inequality turns into
v
t t 2 + 47
2 ≥ 3 − t.
25t 2 + 23
L
.M
25t 4 − 150t 3 + 244t 2 − 138t + 19 ≤ 0,
D
A
PI
(t − 1)2 (5t − 1)(5t − 19) ≤ 0.
M
LY
tion).
O
.M

v v v
7a 7b 7c
W
t t t
+ + ≤ 3.
a + 3b + 3c b + 3c + 3a c + 3a + 3b
W

W
First Solution. Making the substitution

v v v
t 7a t 7b t 7c
x= , y= , z= ,
a + 3b + 3c b + 3c + 3a c + 3a + 3b
we have 

 (x 2 − 7)a + 3x 2 b + 3x 2 c = 0


3 y 2 a + ( y 2 − 7)b + 3 y 2 c = 0 ,


 3z a + 3z b + (z − 7)c = 0

 2 2 2

which involves
x2 − 7 3x 2 3x 2

3 y2 y2 − 7 3 y2 =0 ;

3z 2 3z 2 z2 − 7
that is,
F (x, y, z) = 0,
where X X
F (x, y, z) = 4x 2 y 2 z 2 + 8 x2 y2 + 7 x 2 − 49.
We need to show that F (x, y, z) = 0 involves x + y + z ≤ 3, where x, y, z > 0. To
do this, we use the contradiction method. Assume that x + y + z > 3 and show that
L
F (x, y, z) > 0. Since F (x, y, z) is strictly increasing in each of its arguments, it is enough
.M
to prove that x + y + z = 3 involves F (x, y, z) ≥ 0. Assume that x = max{x, y, z} and
denote
y +z
D
t= , 0 < t ≤ 1 ≤ x.
2
A
We will show that
PI
F (x, y, z) ≥ F (x, t, t) ≥ 0.
M
We have
F (x, y, z) − F (x, t, t) = (8x 2 + 7)( y 2 + z 2 − 2t 2 ) − 4(x 2 + 2)(t 4 − y 2 z 2 )

LY
1
= (8x 2 + 7)( y − z)2 − (x 2 + 2)(t 2 + yz)( y − z)2
O
2
1
≥ (8x 2 + 7)( y − z)2 − 2(x 2 + 2)( y − z)2
.M
2
1
= (4x 2 − 1)( y − z)2 ≥ 0
2
W
and
W
3− x 3− x 1

F (x, t, t) = F x, , = (x − 1)2 (x − 2)2 (x 2 − 6x + 23) ≥ 0.
W
2 2 4
a
The equality holds for a = b = c, and also for
= b = c (or any cyclic permutation).
8
Second Solution. Due to homogeneity, we may assume that a + b + c = 3, when the
inequality becomes v
X t 7a
≤ 3.
9 − 2a
v v v
t 7a t 7b t 7c
x= , y= , z= ,
9 − 2a 9 − 2b 9 − 2c
we need to show that if x, y, z are positive real number such that

X x2 1
= ,
2x + 7 3
2
then
x + y + z ≤ 3.
For the sake of contradiction, assume that x + y + z > 3 and show that F (x, y, z) > 0,
where
X x2 1
F (x, y, z) = − .
2x + 7 3
2
L
Since F (x, y, z) is strictly increasing in each of its arguments, it is enough to prove that
.M
x + y + z = 3 involves F (x, y, z) ≥ 0. This is just the inequality in P 1.30. We give here
another proof. By the Cauchy-Schwarz inequality, we have
D
( x 3/2 )2 ( x 3/2 )2
P P
X x2
≥P = P .
A
2x 2 + 7 x(2x 2 + 7) 2 x 3 + 21

PI
M
X X
3( x 3/2 )2 ≥ 2 x 3 + 21,
LY
which is equivalent to the homogeneous inequality

O
X X 7
x3 + 6 (x y)3/2 ≥ (x + y + z)3 .
9
.M
Replacing x, y, z by x 2 , y 2 , z 2 , we need to prove that G(x, y, z) ≥ 0, where

X X 7
W
G(x, y, z) = x6 + 6 x 3 y 3 − (x 2 + y 2 + z 2 )3 .
9
W
Assume that x = max{x, y, z} and denote

W
v
t y 2 + z2
t= , 0 < t ≤ x.
2
We will show that
G(x, y, z) ≥ G(x, t, t) ≥ 0.
We have
G(x, y, z) − G(x, t, t) = y 6 + z 6 + 6 y 3 z 3 − 8t 4 + 6x 3 ( y 3 + z 3 − 2t 3 ).
Since y 3 + z 3 − 2t 3 ≥ 0, we get
6x 3 ( y 3 + z 3 − 2t 3 ) ≥ 4x 3 ( y 3 + z 3 − 2t 3 ) ≥ ( y 3 + z 3 + 2t 3 )( y 3 + z 3 − 2t 3 ).
Thus,
G(x, y, z) − G(x, t, t) ≥ y 6 + z 6 + 6 y 3 z 3 − 8t 4 + [( y 3 + z 3 )2 − 4t 6 ]
= 2( y 6 + z 6 + 4 y 3 z 3 − 6t 6 ) = 2[( y 2 + z 2 )3 − 3 y 2 z 2 ( y 2 + z 2 ) + 4 y 3 z 3 − 6t 6 ]
= 4(t 6 − 3t 2 y 2 z 2 + 2 y 3 z 3 ) = 4(t 2 − yz)2 (t 2 + 2 yz) ≥ 0.
Also,
7
G(x, t, t) = x 6 + 2t 6 + 6(t 6 + 2x 3 t 3 ) − (x 2 + 2t 2 )3
9
2
= (x − t)2 (x − 2t)2 (x 2 + 6x t + 2t 2 ) ≥ 0.
9
L
D
.M
P 2.83. If a, b, c are positive real numbers such that a + b + c = 3, then
A
Æ
3
Æ
3
Æ
3
p
3
a2 (b2 + c 2 ) + b2 (c 2 + a2 ) + c 2 (a2 + b2 ) ≤ 3 2.
PI (Michael Rozenberg, 2013)
M
LY
X Æ
3
3 X 2 X b 2 + c 2
a2 (b2 + c 2 ) ≤ a(b + c) · .
O
(b + c)2
.M

X b2 + c 2 27
≤ ,
W
(b + c)2 2(a b + bc + ca)2

W
which is equivalent to the homogeneous inequalities

W
X b2 + c 2 p4

− 1 ≤ − 3,
(b + c)2 6q2
X 2bc p4
+ ≥ 3,
(b + c)2 6q2
where
p = a + b + c, q = a b + bc + ca.
According to P 1.61, the following inequality holds
X 2bc p2 9
+ ≥ .
(b + c)2 q 2
9 p2 p4
− + 2 ≥ 3,
2 q 6q
2
p2

−3 ≥ 0.
q
L
.M
D
1 1 1 1 2
+ + ≥ +p .
a+b b+c c+a a+b+c a b + bc + ca
A
M
Solution. Using the notation
LY
O
p2 + q
.M
1 2
≥ +p .
pq − r p q
W
According to P 3.57-(a) in Volume 1, for fixed p and q, the product r = a bc is minimal

when two of a, b, c are equal or one of a, b, c is zero. Therefore, it suffices to prove the
W
inequality for b = c = 1 and for a = 0. For a = 0, the inequality reduces to

W
1 1 2
+ ≥p ,
b c bc
which is obvious. For b = c = 1, the inequality becomes as follows:
1 2 1 2
+ ≥ +p ,
2 a+1 a+2 2a + 1
1 1 2 2
− ≥p − ,
2 a+2 2a + 1 a + 1
p
a 2(a + 1 − 2a + 1)
≥ p ,
2(a + 2) (a + 1) 2a + 1
a 2a2
≥ p p .
2(a + 2) (a + 1) 2a + 1(a + 1 + 2a + 1)
So, we need to show that
1 2a
≥ p p .
2(a + 2) (a + 1) 2a + 1(a + 1 + 2a + 1)
Consider two cases: 0 ≤ a ≤ 1 and a > 1.
Case 1: 0 ≤ a ≤ 1. Since
p p p p p
L
2a + 1(a + 1 + 2a + 1) ≥ 2a + 1( 2a + 1 + 2a + 1) = 2(2a + 1),
.M
1 a
D
≥ ,
2(a + 2) (a + 1)(2a + 1)
A
which is equivalent to 1 − a ≥ 0.
Case 2: a > 1. We will show first that

PI
M
p
(a + 1) 2a + 1 > 3a.
LY
Indeed, by squaring, we get the obvious inequality

O
a3 + a(a − 2)2 + 1 > 0.

.M

W
1 2a
≥ ,
W
2(a + 2) (a + 1)(3a + 2a + 1)
W
which is equivalent to (a−1)2 ≥ 0. The proof is completed. The equality holds for a = 0
and b = c.
P 2.85. If a, b ≥ 1, then
1 1 1 1
p + ≥p +p .
3a b + 1 2 3a + 1 3b + 1
Solution. Using the substitution

2 2
x=p , y=p , x, y ∈ (0, 1],
3a + 1 3b + 1
the desired inequality can be written as
v
t 3
xy ≥ x + y − 1.
x y − x2 − y2 + 4
2 2
Consider the non-trivial case x + y − 1 ≥ 0, and denote
t = x + y − 1, p = x y, 1 ≥ p ≥ t ≥ 0.
L
.M
Since
x 2 + y 2 = (x + y)2 − 2x y = (t + 1)2 − 2p,
D
we need to prove that v
A
t 3
p ≥ t.
p2 + 2p − t 2 − 2t + 3
PI
By squaring, we get the inequality
M
(p − t)[(3 − t 2 )p + t(1 − t)(3 + t)] ≥ 0,
LY
which is clearly true. The equality holds for a = b = 1.

O
.M
P 2.86. Let a, b, c be positive real numbers such that a ≥ 1 ≥ b ≥ c and a bc = 1. Prove

that
W
1 1 1 3
p +p +p ≥ .
3a + 1 3b + 1 3c + 1 2
W

W
Solution. Let b1 = 1/b, b1 ≥ 1. We claim that

1 1 1
p +p ≥ .
3b + 1 3b1 + 1 2
v
1 t b 1
p + ≥ .
3b + 1 b+3 2
Making the substitution
1 1
p = t, ≤ t < 1,
3b + 1 2
the inequality becomes v

t 1 − t2
≥ 1 − t.
1 + 8t 2
By squaring, we get t(1 − t)(1 − 2t)2 ≥ 0, which is clearly true. Similarly, we have
1 1 1
p +p ≥ ,
3c + 1 3c1 + 1 2
where c1 = 1/c, c1 ≥ 1. Using these inequality, it suffices to show that
1 1 1 1
+ ≥p +p ,
L
p
3a + 1 2 3b1 + 1 3c1 + 1
.M
D
1 1 1 1
+ ≥p +p .
A
3b1 c1 + 1 3b1 + 1 3c1 + 1
p
2
PI
From the inequality in the preceding P 2.85, the conclusion follows. The equality holds
for a = b = c = 1.
M
LY
1
O
P 2.87. Let a, b, c be positive real numbers such that a + b + c = 3. If k ≥ p , then

2
.M
(a bc)k (a2 + b2 + c 2 ) ≤ 3.
W
Solution. Since
W
3
a+b+c

abc ≤ = 1,
W
3
p
it suffices to prove the desired inequality for k = 1/ 2. Write the inequality in the
homogeneous form
3k+2
a+b+c

2 2 2
k
(a bc) (a + b + c ) ≤ 3 .
3
According to P 3.57-(a) in Volume 1, for fixed a+b+c and a b+bc+ca, the product a bc is
maximal when two of a, b, c are equal. Therefore, it suffices to prove the homogeneous
inequality for b = c = 1; that is, f (a) ≥ 0, where
f (a) = (3k + 2) ln(a + 2) − (3k + 1) ln 3 − k ln a − ln(a2 + 2).

From
3k + 2 k 2a
f 0 (a) = − − 2
a+2 a a +2
p p
2(a − 1)(ka2 − 2a + 2k) 2(a − 1)(a − 2)2
= = ,
a(a + 2)(a2 + 2) a(a + 2)(a2 + 2)
it follows that f is decreasing on (0, 1] and increasing on [1, ∞); therefore, f (a) ≥
f (1) = 0. This completes the proof. The equality holds for a = b = c.
L
.M
P 2.88. Let p and q be nonnegative real numbers such that p2 ≥ 3q, and let
v v
t 2p − 2w t 2p + w
D
g(p, q) = +2 ,
3 3
A
 s


2p + 2w
+2
s
2p − w
,
PI p2 ≤ 4q
3 3
M
h(p, q) = p p ,
 p + p + q, p2 ≥ 4q
p

LY
where w =
p
p2 − 3q. If a, b, c are nonnegative real numbers such that
O
a + b + c = p, a b + bc + ca = q,
.M
then
p p p
W
(a) a+b+ b+c+ c + a ≥ g(p, q),

W
p + 2w p−w
with equality for a = and b = c = (or any cyclic permutation);
3 3
W
p p p
(b) a + b + b + c + c + a ≤ h(p, q),
p − 2w p+w
with equality for a = and b = c = (or any cyclic permutation) - when
3 3
p2 ≤ 4q, and for a = 0, b + c = p and bc = q (or any cyclic permutation) - when p2 ≥ 4q.
Solution. Consider the non-trivial case p > 0. Since
b + c = p − a,
(a + b)(a + c) = a2 + q
and
p
p Ç Æ
a+b+ a+c = a+p+2 a2 + q,
we get p p p
a+b+ b+c+ c + a = f (a),
where Ç Æ
f (a) = p−a+ a+p+2 a2 + q.
p
From
a2 + q + 2a
p
−1
L
0
f (a) = p + p ,
2 p − a 2 a 2 + q · a + p + 2p a 2 + q
q
.M
it follows that f 0 (a) has the same sign as F (a), where
D
( a2 + q + 2a)2 −2(3a2 − 2pa + q)(a + a2 + q)
p p
−1
A
F (a) = + = .
p − a (a2 + q)(a + p + 2 a2 + q) (p − a)(a2 + q)(a + p + 2 a2 + q)
p p
PI
Therefore, f (a) is increasing on [a1 , a2 ] and decreasing on [0, a1 ] and [a2 , p], where
M
p + p2 − 3q
p p
p − p2 − 3q
a1 = , a2 =
LY
.
3 3
O
(a) From
0 ≤ (b + c)2 − 4bc = (p − a)2 − 4(a2 − pa + q) = −(3a2 − 2pa + 4q − p2 ),

.M
we get a ∈ [0, a4 ], where

p+2
p
p2 − 3q
W
a4 = .
3
W
Since a2 ≤ a4 , f (a) is increasing on [a1 , a2 ] and decreasing on [0, a1 ] ∪ [a2 , a4 ]; there-

fore,
W
f (a) ≥ min{ f (a1 ), f (a4 )}.

min{ f (a1 ), f (a4 )} = g(p, q).
Indeed, from 2a1 + a4 = p and
q Æ q
a12 + q = 2a1 (p − a1 ), 2 a42 + q = a4 + p,
we get
f (a1 ) = f (a4 ) = g(p, q).
(b) Consider the cases 3q ≤ p2 ≤ 4q and p2 ≥ 4q.

Case 1: 3q ≤ p2 ≤ 4q. From
(b + c)2 − 4bc = (p − a)2 − 4(a2 − pa + q) = −(3a2 − 2pa + 4q − p2 ) ≥ 0,
we get a ∈ [a3 , a4 ], where
p+2
p p
p−2 p2 − 3q p2 − 3q
a3 = ≥ 0, a4 = .
3 3
Since a3 ≤ a1 ≤ a2 ≤ a4 , it follows that f (a) is increasing on [a1 , a2 ] and decreasing on
L
[a3 , a1 ] ∪ [a2 , a4 ]. Thus,
.M
f (a) ≤ max{ f (a2 ), f (a3 )}.
To complete the proof, we need to show that
D
max{ f (a2 ), f (a3 )} = h(p, q).
A
Indeed, from 2a2 + a3 = p and PI
M
q Æ q
a22 + q = 2a2 (p − a2 ), 2 a32 + q = a3 + p,
LY
we get
f (a2 ) = f (a3 ) = h(p, q).
O
Case 2: p2 ≥ 4q. Assume that a = min{a, b, c}, a ≤ p/3. From

.M
0 ≤ bc = q − a(b + c) = q − a(p − a) = a2 − pa + q,
we get a ∈ [0, a5 ] ∪ [a6 , p], where

W
p+
p p
W
p − p2 − 4q p2 − 4q
a5 = , a6 = .
2 2
W
Since a6 > p/3, it follows that a ∈ [0, a5 ]. Since a1 ≤ a5 ≤ a2 , f (a) is decreasing on

[0, a1 ] and increasing on [a1 , a5 ]; thus,
f (a) ≤ max{ f (0), f (a5 )}.
It remains to show that max{ f (0), f (a5 )} = h(p, q). Indeed, from a52 + q = pa5 and
r q Æ p
a5 + p − a5 = p+2 pa5 − a52 = p + 2 q,
p p
we get
f (0) = f (a5 ) = h(p, q).
Remark. Note the following particular cases:
(a) If a, b, c are nonnegative real numbers such that
a + b + c = a b + bc + ca = 4,
then p
2(1 + 10) p p p p
p ≤ a + b + b + c + c + a ≤ 2(1 + 2),
3
L
with left equality for a = 8/3 and b = c = 2/3 (or any cyclic permutation), and right
.M
equality for a = 0 and b = c = 2 (or any cyclic permutation).
D
(b) If a, b, c are nonnegative real numbers such that
A
a + b + c = 4, a b + bc + ca = 5,
then
PI
p p
M
p p p p p 10 + 2 7
2+2 3≤ a+ b+ b+c+ c+a ≤ p ,
3
LY
with left equality for a = 2 and b = c = 1 (or any cyclic permutation), and right equality
for a = 2/3 and b = c = 5/3 (or any cyclic permutation).
O
.M
(c) If a, b, c are nonnegative real numbers such that
a + b + c = 11, a b + bc + ca = 7,
W
then
W
p p p p p Æ p
3 6≤ a+b+ b+c+ c+a≤ 11 + 11 + 7,
W
with left equality for a = 31/3 and b = c = 1/3 (or any cyclic permutation), and right
equality for a = 0, b + c = 11 and bc = 7 (or any cyclic permutation).
P 2.89. Let a, b, c, d be positive real numbers such that a2 + b2 + c 2 + d 2 = 1. Prove that

p p p p p p p p
1 − a + 1 − b + 1 − c + 1 − d ≥ a + b + c + d.

First Solution. We can obtain the desired inequality by summing the inequalities
p p p p
1 − a + 1 − b ≥ c + d,
p p p p
1 − c + 1 − d ≥ a + b.
Since p p Æ
4
1 − a + 1 − b ≥ 2 (1 − a)(1 − b)
and v v
p p tc + d 4 c +d
t 2 2
c+ d ≤2 ≤2 ,
2 2
L
the former inequality above holds if
.M
c2 + d 2
(1 − a)(1 − b) ≥ .
D
2
A
Indeed,
PI
2(1 − a)(1 − b) − c 2 − d 2 = 2(1 − a)(1 − b) + a2 + b2 − 1 = (a + b − 1)2 ≥ 0.
M
1
Similarly, we can prove the second inequality. The equality holds for a = b = c = d = .
2
LY
Second Solution. We can obtain the desired inequality by summing the inequalities
O
p p 1 p p 1
1 − a − a ≥ p (1 − 4a2 ), 1− b− b ≥ p (1 − 4b2 ),
.M
2 2 2 2
p p 1 p p 1
1 − c − c ≥ p (1 − 4c 2 ), 1−d − d ≥ p (1 − 4d 2 ).
W
2 2 2 2
To prove the first inequality, we write it as
W
1 − 2a 1
p p ≥ p (1 − 2a)(1 + 2a).
W
1−a+ a 2 2
1
Case 1: 0 < a ≤ . We need to show that
2
p p p
2 2 ≥ (1 + 2a)( 1 − a + a).
p p p p
Since 1 − a + a ≤ 2[(1 − a) + a] = 2, we have
p p p p
2 2 − (1 + 2a)( 1 − a + a) ≥ 2(1 − 2a) ≥ 0.
1
Case 2: ≤ a < 1. We need to show that
2
p p p
2 2 ≤ (1 + 2a)( 1 − a + a).
p
Since 1 + 2a ≥ 2 2a, it suffices to prove that
Æ
1 ≤ a(1 − a) + a.
Indeed,
p
Æ p p p 1 − a (1 − 2a)
1−a− a(1 − a) = 1 − a ( 1 − a − a) = p ≤ 0.
L
p
1−a+ a
D
.M
A
P 2.90. Let a, b, c, d be positive real numbers. Prove that
p
A + 2 ≥ B + 4,
PI
M
where
1 1 1 1

LY
A = (a + b + c + d) + + + − 16,
a b c d
O
1 1 1 1

2 2 2 2
B = (a + b + c + d ) 2 + 2 + 2 + 2 − 16.
a b c d
.M

W
A2 + 4A ≥ B.
W
Let us denote
W
x y z x2 y 2 z2
f (x, y, z) = + + − 3, F (x, y, z) = + + 2 − 3,
y z x y 2 z2 x
where x, y, z > 0. By the AM-GM inequality, it follows that f (x, y, z) ≥ 0 and F (x, y, z) ≥
0. We can check that
A = f (a, b, c) + f (b, a, d) + f (c, d, a) + f (d, c, b)

= f (c, b, a) + f (d, a, b) + f (a, d, c) + f (b, c, d)
and
B = F (a, b, c) + F (b, a, d) + F (c, d, a) + F (d, c, b).
Since
F (x, y, z) = [ f (x, y, z) + 3]2 − 2[ f (z, y, x) + 3] − 3

= f 2 (x, y, z) + 6 f (x, y, z) − 2 f (z, y, x),
we get
B = f 2 (a, b, c) + f 2 (b, a, d) + f 2 (c, d, a) + f 2 (d, c, b) + 4A.
Therefore,
A2 + 4A − B = [ f (a, b, c) + f (b, a, d) + f (c, d, a) + f (d, c, b)]2

− f 2 (a, b, c) − f 2 (b, a, d) − f 2 (c, d, a) − f 2 (d, c, b) ≥ 0.
L
.M
The equality holds for a = b = c = d.
D
A
P 2.91. Let a1 , a2 , . . . , an be nonnegative real numbers such that a1 + a2 + · · · + an = 1.
PI
Prove that
3a1 + 1 + 3a2 + 1 + · · · + 3an + 1 ≥ n + 1.
p p p
M
LY
First Solution. Without loss of generality, assume that a1 = max{a1 , a2 , · · · , an }. Write

the inequality as follows:
O
( 3a1 + 1 − 2) + ( 3a2 + 1 − 1) + · · · + ( 3an + 1 − 1) ≥ 0,

p p p
.M
a1 − 1 a2 an
+p + ··· + p ≥ 0,
3a1 + 1 + 2 3a2 + 1 + 1 3an + 1 + 1
p
W
a2 an a2 + · · · + an
+ ··· + p ≥p
W
,
3a2 + 1 + 1 3an + 1 + 1 3a1 + 1 + 2
p
W

1 1 1 1
a2 p −p + · · · + an p −p ≥ 0.
3a2 + 1 + 1 3a1 + 1 + 2 3an + 1 + 1 3a1 + 1 + 2
The last inequality is clearly true. The equality holds for a1 = 1 and a2 = · · · = an = 0
Second Solution. We use the induction method. For n = 1, the inequality is an equality.
We claim that Æ
3a1 + 1 + 3an + 1 ≥ 3(a1 + an ) + 1 + 1.
p p
By squaring, this becomes

Æ Æ
(3a1 + 1)(an + 1) ≥ 3(a1 + an ) + 1,
which is equivalent to a1 an ≥ 0. Thus, to prove the original inequality, it suffices to

show that Æ
3(a1 + an ) + 1 + 3a2 + 1 + · · · + 3an−1 + 1 ≥ n.
p p
Using the substitution b1 = a1 + an and b2 = a2 , · · · , bn−1 = an−1 , this inequality be-

comes Æ Æ Æ
3b1 + 1 + 3b2 + 1 + · · · + 3bn−1 + 1 ≥ n
for b1 + b2 + · · · + bn−1 = 1. Clearly, this is true by the induction hypothesis.
L
P 2.92. Let 0 ≤ a < b and a1 , a2 , . . . , an ∈ [a, b]. Prove that
.M
p p 2
a1 + a2 · · · + an − n n a1 a2 · · · an ≤ (n − 1)
p
b− a .
D
A
Solution. Based on Lemma below, it suffices to consider that PI

M
a1 = · · · = ak = a, ak+1 = · · · = an = b,
LY
where k ∈ {1, 2, · · · , n − 1}. The original inequality becomes

O
k n−k p
(n − k − 1)a + (k − 1)b + na n b n ≥ (2n − 2) a b.
.M
Clearly, this inequality follows by the weighted AM-GM inequality. For n ≥ 3, the equal-
ity holds when a = 0, one of ai is also 0 and the other ai are equal to b.
W
Lemma. Let 0 ≤ a < b and a1 , a2 , . . . , an ∈ [a, b]. Then, the expression

W
a1 + a2 · · · + an − n n a1 a2 · · · an
p
W
is maximal when a1 , a2 , . . . , an ∈ {a, b}.

Proof. We use the contradiction method. Consider that a2 , . . . , an are fixed and define
the function
f (a1 ) = a1 + a2 + · · · + an − n n a1 a2 · · · an .
p
For the sake of contradiction, assume that there exits a1 ∈ (a, b) such that
p f (a1 ) > f (a)
p p n
and f (a1 ) > f (b). Let us denote x i = n ai for all i, c = n a and d = b (c < x 1 < d).
From
f (a1 ) − f (a) = x 1n − c n − n(x 1 − c)x 2 · · · x n

= (x 1 − c)(x 1n−1 + x 1n−2 c + · · · + c n−1 − nx 2 · · · x n ),
we get
x 1n−1 + x 1n−2 c + · · · + c n−1 > nx 2 · · · x n . (*)
Analogously, from
f (a1 ) − f (b) = x 1n − d n − n(x 1 − d)x 2 · · · x n

= (x 1 − d)(x 1n−1 + x 1n−2 d + · · · + d n−1 − nx 2 · · · x n ),
we get
L
nx 2 · · · x n > x 1n−1 + x 1n−2 d + · · · + d n−1 . (**)
.M
Summing up (*) and (**) yields
D
x 1n−1 + x 1n−2 c + · · · + c n−1 > x 1n−1 + x 1n−2 d + · · · + d n−1 ,
A
which is clearly false.
PI
M
LY
1 1 1
O
+p + ··· + p ≥ 1.
1 + (n2 1 + (n2 − 1)a2 1 + (n2 − 1)an
p
− 1)a1
.M
Solution. For the sake of contradiction, assume that

W
1 1 1
+p + ··· + p < 1.
W
1 + (n2 1 + (n2 1 + (n2 − 1)an

p
− 1)a1 − 1)a2
W
It suffices to show that a1 a2 · · · an > 1. Let
1
xi = p , 0 < x i < 1, i = 1, 2, · · · , n.
1 + (n2 − 1)ai
1 − x i2
Since ai = for all i, we need to show that
(n2 − 1)x i2
x1 + x2 + · · · + x n < 1
implies
(1 − x 12 )(1 − x 22 ) · · · (1 − x n2 ) > (n2 − 1)n x 12 x 22 · · · x n2 .
Using the AM-GM inequality gives

Y Y hX 2 i Y
(1 − x 12 ) > x 1 − x 12 = (x 2 + · · · + x n )(2x 1 + x 2 + · · · + x n )
Y p q
≥ (n2 − 1)n x 12 x 2 · · · x n = (n2 − 1)n x 12 x 22 · · · x n2 .
n+1
n−1
x2 · · · x n ·
L
.M
n
X 1 1
≥ .
1+ 1 + 4n(n − 1)ai
p
i=1
2
D
A
Xn p
1 + 4n(n − 1)ai − 1
PI
≥ 2n(n − 1),
M
i=1
a1
LY
v
n u
X
t1 4n(n − 1) X 1
+ ≥ 2n(n − 1) + .
O
i=1
a12 a1 a1
.M

v
X u 1
u
4n(n − 1) 1 4n(n − 1) X 1
W
2 2
t
2
+ 2
+ ≥ 2n (n − 1) + .
1≤i< j≤n
a i
a i a j
a j 1≤i< j≤n
a i a j
W
The Cauchy-Schwarz inequality gives

W
v
u
u 1 4n(n − 1) 1 4n(n − 1) 1 4n(n − 1)
t
2
+ 2
+ ≥ + p .
ai ai aj aj ai a j ai a j

X 1 n(n − 1)
≥ ,
ai a j 2
p
1≤i< j≤n
which follows immediately from the AM-GM inequality. The equality holds for a1 =
a2 = · · · = an = 1.
Second Solution. For the sake of contradiction, assume that

n
X 1 1
< .
1+ 1 + 4n(n − 1)ai
p
i=1
2
It suffices to show that a1 a2 · · · an > 1. Using the substitution

xi 1
= , i = 1, 2, · · · , n,
2n 1 + 1 + 4n(n − 1)ai
p
which yields
n − xi
ai = , 0 < x i < n, i = 1, 2, · · · , n,
L
(n − 1)x i2
.M
x1 + x2 + · · · + x n < n
D
implies
A
(n − x 1 )(n − x 2 ) · · · (n − x n ) > (n − 1)n x 12 x 22 · · · x n2 .
PI
M
x + x + · · · + x n
1 2 n
x1 x2 · · · x n ≤ <1
n
LY
and
O
v
t x1 x2 · · · x n
n − x i > (x 1 + x 2 + · · · + x n ) − x i ≥ (n − 1) n−1
, i = 1, 2, · · · , n.
.M
xi
Therefore, we get
W
(n − x 1 )(n − x 2 ) · · · (n − x n ) > (n − 1)n x 1 x 2 · · · x n > (n − 1)n x 12 x 22 · · · x n2 .

W
W
P 2.95. If f is a convex function on a real interval I and a1 , a2 , . . . , an ∈ I, then

a + a + ··· + a
1 2 n
f (a1 ) + f (a2 ) + · · · + f (an ) + n(n − 2) f ≥
n
≥ (n − 1)[ f (b1 ) + f (b2 ) + · · · + f (bn )],
where
1 X
bi = aj, i = 1, 2, · · · , n.
n − 1 j6=i
(Tiberiu Popoviciu, 1965)

Solution. Without loss of generality, we may assume that n ≥ 3 and a1 ≤ a2 ≤ · · · ≤ an .

There is an integer m, 1 ≤ m ≤ n − 1, such that
a1 ≤ · · · ≤ am ≤ a ≤ am+1 ≤ · · · ≤ an ,
b1 ≥ · · · ≥ bm ≥ a ≥ bm+1 ≥ · · · ≥ bn .
We can get the desired inequality by summing the following two inequalities:
f (a1 ) + f (a2 ) + · · · + f (am ) + n(n − m − 1) f (a) ≥
≥ (n − 1)[ f (bm+1 ) + f (bm+2 ) + · · · + f (bn )], (*)
f (am+1 ) + f (am+2 ) + · · · + f (an ) + n(m − 1) f (a) ≥
L
≥ (n − 1)[ f (b1 ) + f (b2 ) + · · · + f (bm )]. (**)
.M
In order to prove (*), we apply Jensen’s inequality to get
D
f (a1 ) + f (a2 ) + · · · + f (am ) + (n − m − 1) f (a) ≥ (n − 1) f (b),
A
where
a1 + a2 + · · · + am + (n − m − 1)a
PI
b= .
n−1
M
LY
(n − m − 1) f (a) + f (b) ≥ f (bm+1 ) + f (bm+2 ) + · · · + f (bn ).

Since
O
a ≥ bm+1 ≥ bm+2 ≥ · · · ≥ bn , (n − m − 1)a + b = bm+1 + bm+2 + · · · + bn ,

.M
−
→ −
→
we see that A n−m = (a, · · · , a, b) majorizes B n−m = (bm+1 , bm+2 , · · · , bn ). Therefore,
the inequality is a consequence of Karamata’s inequality.
W
Similarly, we can prove the inequality (**), by summing Jensen’s inequality

W
f (am+1 ) + f (am+2 ) + · · · + f (an ) + (m − 1) f (a)

≥ f (c)
n−1
W
and the inequality

f (c) + (m − 1) f (a) ≥ f (b1 ) + f (b2 ) + · · · + f (bm ),
where
am+1 + am+2 + · · · + an + (m − 1)a
c= .
n−1
The last inequality follows by Karamata’s inequality, because
b1 ≥ b2 ≥ · · · ≥ bm ≥ a, c + (m − 1)a = b1 + b2 + · · · + bm ,
−
→ −
→
therefore C m = (c, a, · · · , a) majorizes D m = (b1 , b2 , · · · , bm ).
P 2.96. Let a1 , a2 , . . . , an (n ≥ 3) be positive real numbers such that a1 a2 · · · an = 1. Prove

that
n
X 1 1
≤ .
i=1 n − 1 + (n − 1)2 + 4nai
p
2
Solution. Use the contradiction method. Assume that

n
X 1 1
> ,
n−1+ (n − 1)2 + 4nai
p
i=1
2
and show that a1 a2 · · · an < 1. Using the substitution
L
.M
1 n − ci
= , i = 1, 2, · · · , n,
n−1+ (n − 1)2 + 4nai 2n(n − 1)
p
D
which involves
(n − 1)2 ci
A
ai = , 0 < ci < n, i = 1, 2, · · · , n,
(n − ci )2
PI
c1 + c2 + · · · + cn < n
M
implies
LY
n − c 2 n − c 2 n − c 2
1 2 n
··· > c1 c2 · · · cn .
n−1 n−1 n−1
O

.M
c1 + c2 + · · · + cn = n
implies
W
n − c 2 n − c 2 n − c 2
1 2 n
···
≥ c1 c2 · · · cn .
n−1 n−1 n−1
W
Popoviciu’s inequality (see the preceding P 2.95) applied to the convex function f (x) =
W
− ln x, x > 0 gives
n − c n−1 n − c n−1 n − c n−1 c + c + · · · + c n(n−2)
1 2 n 1 2 n
··· ≥ c1 c2 · · · cn ,
n−1 n−1 n−1 n
n − c n−1 n − c n−1 n − c n−1
1 2 n
··· ≥ c1 c2 · · · cn ,
n−1 n−1 n−1
n − c 2 n − c 2 n − c 2 2
1 2 n
··· ≥ (c1 c2 · · · cn ) n−1 .
n−1 n−1 n−1
2
(c1 c2 · · · cn ) n−1 ≥ c1 c2 · · · cn .
For n = 3, this inequality is an equality, while for n ≥ 4, it is equivalent to c1 c2 · · · cn ≤ 1.

Indeed, by the AM-GM inequality, we have
c + c + · · · + c n
1 2 n
c1 c2 · · · cn ≤ = 1.
n
L
P 2.97. If a1 , a2 , . . . , an are positive real numbers such that a1 a2 · · · an = 1, then
.M
v
t a1 + a22 + · · · + an2
u 2
a1 + a2 + · · · + an ≥ n − 1 + .
D
n
A
PI
M
v P
a1 + a2 + · · · + an t 2 1≤i< j≤n ai a j
u
LY
a= , b= ,
n n(n − 1)
O
where b ≥ 1 (by the AM-GM inequality). We need to show that

.M
v
t n2 a2 − n(n − 1)b2
na − n + 1 ≥ .
n
W

W
(n − 1)[n(a − 1)2 + b2 − 1] ≥ 0,
W
which is clearly true. The equality holds for a1 = a2 = · · · = an = 1.
P 2.98. If a1 , a2 , . . . , an are positive real numbers such that a1 a2 · · · an = 1, then

q Æ
(n − 1)(a12 + a22 + · · · + an2 ) + n − n(n − 1) ≥ a1 + a2 + · · · + an .

Solution. We use the induction method. For n = 2, the inequality is equivalent to the
obvious inequality
1
a1 + ≥ 2.
a1
Assume now that the inequality holds for n − 1 numbers, n ≥ 3, and prove that it holds
also for n numbers. Let a1 = min{a1 , a2 , . . . , an }, and denote
a2 + a3 + · · · + an
x= y=
p
, n−1
a2 a3 · · · an ,
n−1
q Æ
f (a1 , a2 , . . . , an ) = (n − 1)(a12 + a22 + · · · + an2 ) + n − n(n − 1) − (a1 + a2 + · · · + an ).
L
By the AM-GM inequality, we have x ≥ y. We will show that
.M
f (a1 , a2 , . . . , an ) ≥ f (a1 , y, · · · , y) ≥ 0. (*)
D
Write the left inequality as
A
q q p
a12 + a22 + · · · + an2 − a12 + (n − 1) y 2 ≥ n − 1 (x − y).
PI
To prove this inequality, we use the induction hypothesis, written in the homogeneous
M
form
LY
q Æ
(n − 2)(a22 + a32 + · · · + an2 ) + n − 1 − (n − 1)(n − 2) y ≥ (n − 1)x,
O
a22 + · · · + an2 ≥ (n − 1)A2 ,

.M
where v
tn − 1
A = k x − (k − 1) y, k= .
n−2
W
So, we need to prove that

W
q q p
a12 + (n − 1)A2 − a12 + (n − 1) y 2 ≥ n − 1 (x − y).
W
A2 − y 2 x−y
q q ≥p .
a1 + (n − 1)A2 + a1 + (n − 1) y 2
2 2 n−1
Since x ≥ y and
A2 − y 2 = k(x − y)[k x − (k − 2) y],
k[k x − (k − 2) y] 1
q q ≥p .
a12 + (n − 1)A2 + a1 + (n − 1) y 2
2 n−1
In addition, since a1 ≤ y, it suffices to show that
k[k x − (k − 2) y] 1
p ≥p .
y 2 + (n − 1)A2 + n y
p
n−1
We claim that Æ p
y 2 + (n − 1)A2 ≤ k n − 1 [k x − (k − 1) y] .
If this inequality is true, then it is enough to prove that
k[k x − (k − 2) y] 1
p p ≥p .
k n − 1[k x − (k − 1) y] + ny n−1
L
.M
k[k x − (k − 2) y]
s
n
D
≥ 1, m= .
k[k x − (k − 1) y] + m y n−1
A
Since m < k, we have PI
k[k x − (k − 2) y] k[k x − (k − 2) y]
M
> = 1.
k[k x − (k − 1) y] + m y k[k x − (k − 1) y] + k y
LY
Using the relation a1 y n−1 = 1, we can write the right inequality in (*) as
O
Æ
(n − 1)[(n − 1) y 2n + 1] ≥ (n − 1) y n − t y n−1 + 1,
.M
where Æ
t = n− n(n − 1).
W

W
(n − 1)[(n − 1) y 2n + 1] ≥ [(n − 1) y n − t y n−1 + 1]2 ,

W
that is,
2(n − 1)t y 2n−1 − t 2 y 2n−2 − 2(n − 1) y n + 2t k y n−1 + n − 2 ≥ 0.
Since
t 2 = n(2t − 1),
we can write this inequality in the form
2t y n−1 B + C ≥ 0,
where
B = (n − 1) y n − n y n−1 + 1, C = n y 2n−2 − 2(n − 1) y n + n − 2.
This inequality is true if b ≥ 0 and C ≥ 0. Indeed, by the AM-GM inequality, we have

Æ
n
(n − 1) y n + 1 ≥ n y (n−1)n · 1 = n y n−1
and Æ
2n−2
n y 2n−2 + n − 2 ≥ (2n − 2) y n(2n−2) · 1n−2 = 2(n − 1) y n .
The proof is completed. The equality holds for a1 = a2 = · · · = an = 1.
P 2.99. Let a1 , a2 , . . . , an (n ≥ 3) be positive real numbers such that a1 a2 · · · an = 1. If
L
.M
2n − 1
0<p≤ ,
(n − 1)2
D
then
1 1 1 n
+p +p
A
≤p .
1 + pa1 1 + pa2 1 + pan 1+p
p
PI
(Vasile Cîrtoaje and Gabriel Dospinescu, 2006)
M
Solution. We use the contradiction method. Assume that the reverse inequality holds,
LY
namely
1 1 1 n
+p +p >p ,
O
1 + pa1 1 + pa2 1 + pan 1+p

p
and show that this implies a1 a2 · · · an < 1. Using the substitution

.M
1+p
p
p
1 + pai = , 0 < x i < p + 1, i = 1, 2, · · · , n,
p
W
xi
we need to show that x 1 + x 2 + · · · + x n > n yields
W
1+p 1+p 1+p

W
−1 − 1 ··· − 1 < pn.

x 12 x 22 x n2
Clearly, it suffices to prove that
x1 + x2 + · · · + x n = n
yields
1+p 1+p 1+p

−1 − 1 ··· − 1 ≤ pn.
x 12 x 22 x n2
Denoting
p n
1 + p = q, 1<q≤ ,
n−1
(q2 − x 12 )(q2 − x 22 ) · · · (q2 − x n2 ) ≤ (q2 − 1)n (x 1 x 2 · · · x n )2 (*)
for all x i ∈ (0, q) satisfying x 1 + x 2 + · · · + x n = n. Applying Popoviciu’s inequality (see

P 2.95) to the convex function
n
f (x) = − ln − x , 0 < x < 1,
n−1
we get
(x 1 x 2 · · · x n )n−1 ≥ [n − (n − 1)x 1 ][n − (n − 1)x 2 ] · · · [n − (n − 1)x n ]. (**)
L
.M
On the other hand, Jensen’s inequality applied to the convex function
n − (n − 1)x
f (x) = ln
D
q−x
A
yields
[n − (n − 1)x 1 ][n − (n − 1)x 2 ] · · · [n − (n − 1)x n ]
PI 1
≥ .
(q − x 1 )(q − x 2 ) · · · (q − x n ) (q − 1)n
M
Multiplying this inequality and (**) gives
LY
(q − x 1 )(q − x 2 ) · · · (q − x n )
(x 1 x 2 · · · x n )n−1 ≥ .
(q − 1)n
O
Therefore, in order to prove (*), we still have to show that

.M
(x 1 x 2 · · · x n )n−3 (q + x 1 )(q + x 2 ) · · · (q + x n ) ≤ (q + 1)n .
This is true because, by the AM-GM inequality, we have

W
x + x + · · · + x n
1 2 n
x1 x2 · · · x n ≤ =1
W
n
W
and x 1 + x 2 + · · · + x n n
(q + x 1 )(q + x 2 ) · · · (q + x n ) ≤ q + = (q + 1)n .
n
The equality occurs for a1 = a2 = · · · = an = 1.
2n − 1
Remark. For p = , we get the following inequality
(n − 1)2
n
X 1
≤ 1,
(n − 1)2 + (2n − 1)ai
p
i=1
which holds for all positive numbers a1 , a2 , . . . , an (n ≥ 3) satisfying a1 a2 · · · an = 1.

P 2.100. If a1 , a2 , . . . , an (n ≥ 3) are positive real numbers such that a1 a2 · · · an = 1, then
n Ç
n
2
X X
(n − 1)2 ai4 + 2n − 1 ≥ ai ;
i=1 i=1
n
n
2
X q X
ai (n − 1)2 ai2 + 2n − 1 ≥ ai .
i=1 i=1
L
.M
Solution. According to the preceding P 2.99, the following inequality holds for any real
m:
n
X 1
D
≤ 1.
(n − 1)2 + (2n − 1)aim
Æ
i=1
A
On the other hand, by the Cauchy-Schwarz inequality, we have
n
n
PI n
2
1
M
X X q X
ai2 (n − 1)2 + (2n − 1)aim ≥ ai ,
(n − 1)2 + (2n − 1)aim
Æ
i=1 i=1 i=1
LY
hence 2
O
n n

X q X
ai2 (n − 1)2 + (2n − 1)aim ≥ ai .
.M
i=1 i=1
Setting m = −4 and m = −2 gives the desired inequalities. The equality occurs for
a1 = a2 = · · · = an = 1.
W
W
W
P 2.101. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an ≥ 1. If k > 1,

then
X a1k
≥ 1.
a1k + a2 + · · · + an
p
First Solution. Let us denote r = n a1 a2 · · · an and bi = ai /r for i = 1, 2, · · · , n. Note
that r ≥ 1 and b1 b2 · · · bn = 1. The desired inequality becomes
X b1k
≥ 1,
b1k + (b2 + · · · + bn )/r k−1
and we see that it suffices to prove it for r = 1; that is, for a1 a2 · · · an = 1. On this
hypothesis, we will show that there exists a positive number p such that
p
a1k a1
≥ p p p.
a1k + a2 + · · · + an a1 + a2 + · · · + an
If this is true, by adding this inequality and the analogous inequalities for a2 , . . . , an , we
get the desired inequality. Write the claimed inequality as
p
a2 + · · · + anp ≥ (a2 · · · an )k−p (a2 + · · · + an ).
For
L
(n − 1)k + 1
p= , p > 1,
.M
n
this inequality turns into the homogeneous inequality
D
p−1
p
a2 + · · · + anp ≥ (a2 · · · an ) n−1 (a2 + · · · + an ).
A
Based on the AM-GM inequality PI
a + · · · + a p−1
p−1
M
2 n
(a2 · · · an ) n−1 ≤ ,
n−1
LY
we only need to show that

p p
a2 + · · · + a n
O
a + · · · + a p
2 n
≥ ,
n−1 n−1
.M
which is just Jensen’s inequality applied to the convex function f (x) = x p . The equality
holds for a1 = a2 = · · · = an = 1.
W

W
2
P k+1 2 k+1
a a1 + 2 1≤i< j≤n (a1 a2 ) 2
P k+1 P
a1k 1
W
X
≥P = P k+1 .
a1k + a2 + · · · + an a1 (a1k + a2 + · · · + an ) a1 + 2 1≤i< j≤n a1 a2
P

X k+1
X
(a1 a2 ) 2 ≥ a1 a2 .
1≤i< j≤n 1≤i< j≤n
k+1
Jensen’s inequality applied to the convex function f (x) = x 2 yields
P k+1
X k+1 n(n − 1) 2 1≤i< j≤n a1 a2 2
(a1 a2 ) 2 ≥
1≤i< j≤n
2 n(n − 1)
On the other hand, by the AM-GM inequality, we get

2 X 2
a1 a2 ≥ (a1 a2 · · · an ) n ≥ 1.
n(n − 1) 1≤i< j≤n
Therefore,
P k+1 P
2 1≤i< j≤n a1 a2 2 2 1≤i< j≤n a1 a2
≥ ,
n(n − 1) n(n − 1)
hence P
X k+1 n(n − 1) 2 1≤i< j≤n a1 a2 X
(a1 a2 ) 2 ≥ = a1 a2 .
L
1≤i< j≤n
2 n(n − 1) 1≤i< j≤n
D
.M
P 2.102. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an ≥ 1. If
A
−2 PI
≤ k < 1,
n−2
M
then
X a1k
LY
≤ 1.
a1k + a2 + · · · + an
O

p
Solution. Let us denote r = n a1 a2 · · · an and bi = ai /r for i = 1, 2, · · · , n. Notice that
.M
r ≥ 1 and b1 b2 · · · bn = 1. The desired inequality becomes
b1k
W
X
≤ 1,
b1k + (b2 + · · · + bn )r 1−k
W
W
hypothesis, we will prove the desired inequality by summing the inequality

p
a1k a1
≤ p p p
a1k + a2 + · · · + an a1 + a2 + · · · + an
and the analogous inequalities for a2 , . . . , an , where

(n − 1)k + 1 −1
p= , ≤ p < 1.
n n−2
1−p
p
a2 + · · · + an ≥ (a2 · · · an ) n−1 (a2 + · · · + anp ). (*)
To prove it, we use the weighted AM-GM inequality
1 + (n − 2)p n − 1 1+(n−2)p 1−p

a2 + a3 + · · · + an ≥ a2 n−1 (a3 · · · an ) n−1 ,
1−p 1−p
1 + (n − 2)p n−1 p 1−p

a2 + a3 + · · · + an ≥ a2 (a2 a3 · · · an ) n−1 .
1−p 1−p
Adding this inequality and the analogous inequalities for a3 , · · · , an yields the inequality
(*). Thus, the proof is completed. The equality holds for a1 = a2 = · · · = an = 1.
L
.M
P 2.103. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an ≥ 1. If k > 1,
D
then
A
X a1
≤ 1.
a1 + a2 + · · · + an
k
M
1
Solution. We consider two cases: 1 < k ≤ n + 1 and k ≥ n − .
LY
n−1
Case 1: 1 < k ≤ n + 1. By the AM-GM inequality, the hypothesis a1 a2 · · · an ≥ 1 yields
O
a1 + a2 + · · · + an ≥ n. Therefore, it suffices to prove that the desired inequality holds

for a1 + a2 + · · · + an ≥ n. Actually, we only need to consider that a1 + a2 + · · · + an = n.
.M
Indeed, if we denote p = (a1 + a2 + · · · + an )/n and bi = ai /p for i = 1, 2, · · · , n, then

the desired inequality becomes
W
X b1
≤ 1, p ≥ 1.
p k−1 b1k + b2 + · · · + bn
W
Clearly, it suffices to consider the case p = 1; that is, a1 + a2 + · · · + an = n. On this

W
hypothesis, we can rewrite the desired inequality as

X a1
≤ 1.
a1 − a1 + n
k
Since k > 1, by Bernoulli’s inequality, we have
a1k − a1 + n ≥ 1 + k(a1 − 1) − a1 + n = n − k + 1 + (k − 1)a1 > 0.

X a1
≤ 1,
n − k + 1 + (k − 1)a1
X 1
≥ 1.
n − k + 1 + (k − 1)a1
This inequality follows immediately from the AM-HM inequality
X X 1
x1 ≥ n2 ,
x1
for x i = n − k + 1 + (k − 1)ai , i = 1, 2, · · · , n.
1 p
Case 2: k ≥ n − . Let us denote r = n a1 a2 · · · an and bi = ai /r for i = 1, 2, · · · , n.
n−1
Note that r ≥ 1 and b1 b2 · · · bn = 1. The desired inequality becomes
L
.M
X b1
≤ 1,
b1k r k−1 + b2 + · · · + b n
D
A
hypothesis, it suffices to show that
(n − 1)a1
PI
p
a1
+ ≤1
M
p p p
a1k + a2 + · · · + an a1 + a2 + · · · + an
LY
for a suitable real p; then, adding this inequality and the analogous inequalities for
p
a2 , · · · an yields the desired inequality. Let us denote t = n−1 a2 · · · an . By the AM-GM
O
inequality, we have
p
a2 + · · · an ≥ (n − 1)t, a2 + · · · + anp ≥ (n − 1)t p .
.M

W
p
(n − 1)a1 a1
+ p ≤ 1.
W
a1k + (n − 1)t a1 + (n − 1)t p

W
Since a1 = 1/t n−1 , this inequality is equivalent to
(n − 1)t n+q − (n − 1)t q − t q−np + 1 ≥ 0,
where
q = (n − 1)(k − 1).
Choosing
(n − 1)(k − n − 1)
p= ,
n
the inequality becomes as follows
(n − 1)t n+q − (n − 1)t q − t n(n−1) + 1 ≥ 0,

2 2
−2n −3n
(n − 1)t q (t n − 1) − (t n − 1)(t n + tn + · · · + 1) ≥ 0,
2 2
−2n −3n
(t n − 1)[(t q − t n ) + (t q − t n ) + · · · + (t q − 1)] ≥ 0.
1
The last inequality is clearly true for q ≥ n2 − 2n; that is, for k ≥ n − . Thus, the
n−1
proof is completed. The equality holds for a1 = a2 = · · · = an = 1.
P 2.104. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an ≥ 1. If
L
2
−1 − ≤ k < 1,
.M
n−2
then X a1
D
≥ 1.
a1k + a2 + · · · + an
A
p
PI
Solution. Let us denote r = n a1 a2 · · · an and bi = ai /r for i = 1, 2, · · · , n. Note that
M
r ≥ 1 and b1 b2 · · · bn = 1. The desired inequality becomes
LY
X b1
≥ 1,
b1k /r 1−k + b2 + · · · + bn
O
.M
hypothesis, by the Cauchy-Schwarz inequality, we have
( a1 )2 ( a1 )2
P P
X a1
= P
W
≥P P .
a1k + a2 + · · · + an a1 (a1k + a2 + · · · + an ) ( a1 )2 + a11+k − a12
P
W
Thus, we still have to show that

W
X X
a12 ≥ a11+k .
Case 1: −1 ≤ k < 1. Using Chebyshev’s inequality and the AM-GM inequality yields
X 1 X 1−k X 1+k X X
a12 ≥ a1 a1 ≥ (a1 a2 · · · an )(1−k)/n a11+k = a11+k .
n
2
Case 2: −1 − ≤ k < −1. It is convenient to replace the numbers a1 , a2 , · · · , an
n−1
(n−1)/2 (n−1)/2
by a1 , a2 , · · · , an(n−1)/2 , respectively. So, we need to show that a1 a2 · · · an = 1
involves X X
q
a1n−1 ≥ a1 ,
where
(n − 1)(1 + k)
q= , −1 ≤ q < 0.
2
By the AM-GM inequality, we get
X 1 X n−1 X X 1
a1n−1 = (a2 + · · · + ann−1 ) ≥ a2 · · · an = .
n−1 a1
Thus, it suffice to show that

X 1 X
q
≥ a1 .
a1
L
By Chebyshev’s inequality and the AM-GM inequality, we have
.M
X 1 1 X −1−q X q X X
q q
≥ a1 a1 ≥ (a1 a2 · · · an )−(1+q)/n a1 = a1 .
a1 n
D
A
Thus, the proof is completed. The equality holds for a1 = a2 = · · · = an = 1.
PI
M
P 2.105. Let a1 , a2 , . . . , an be positive real numbers such that a1 a2 · · · an = 1. If k ≥ 0,
LY
then X 1
O
≤ 1.
a1 + a2 + · · · + an
k
.M
Solution. Consider two cases: 0 ≤ k ≤ 1 and k ≥ 1.

W
Case 1: 0 ≤ k ≤ 1. By the Cauchy-Schwarz inequality and the AM-GM inequality, we

W
have
1 a11−k + 1 + · · · + 1
≤ p
W
p p
a1k + a2 + · · · + an ( a1 + a2 + · · · + an )2
a11−k + n − 1 a11−k + n − 1
=P ≤P ,
a1 + 2 1≤i< j≤n ai a j a1 + n(n − 1)
P p
hence
a11−k + n(n − 1)
P
X 1
≤ .
a1 + n(n − 1)
P
a1k + a2 + · · · + an
X X
a11−k ≤ a1 .
Indeed, by Chebyshev’s inequality and the AM-GM inequality, we have

X X 1 X k X 1−k X X
a1 = a1k · a11−k ≥ a1 a1 ≥ (a1 a2 · · · an )k/n a11−k = a11−k .
n
Case 2: k ≥ 1. Write the inequality as
X p
n−1 a1
+ p p p − 1 ≤ 0,
a1k + a2 + · · · + an a1 + a2 + · · · + an
L
where
(n − 1)(k − 1)
.M
p= ≥ 0.
n
To complete the proof, it suffices to show that
D
p
A
n−1 a1
≤1− p p p.
a1k + a2 + · · · + an a1 + a2 + · · · + an
PI
Let
M
x= x > 0.
p
n−1
a1 ,
LY

O
n−1 n−1
a2 + · · · + an ≥ (n − 1) n−1 a2 · · · an = n−1
p =
p
a1 x
.M
and
p
Æ n−1 n−1
a2 + · · · + anp ≥ (a2 · · · an ) p = q p =
n−1
.
W
n−1
a1 xp
W

W
n−1 x (n−1)p
n−1
≤1− n−1
,
x (n−1)k + x x (n−1)p + xp
x 1
≤ ,
x (n−1)k+1 +n−1 x np + n − 1
(n−1)k+1
x − x np+1 − (n − 1)(x − 1) ≥ 0,
x q (x n−1 − 1) − (n − 1)(x − 1) ≥ 0,
where
q = (n − 1)(k − 1) + 1 ≥ 1.
x q (x n−1 − 1) − (n − 1)(x − 1) = (x q − 1)(x n−1 − 1) + (x n−1 − 1) − (n − 1)(x − 1),
and
(x q − 1)(x n−1 − 1) ≥ 0,
(x n−1 − 1) − (n − 1)(x − 1) = (x − 1)[(x n−2 − 1) + (x n−3 − 1) + · · · + (x − 1)] ≥ 0.
This completes the proof. The equality holds for a1 = a2 = · · · = an = 1.
L
P 2.106. Let a1 , a2 , . . . , an be nonnegative real numbers such that a1 + a2 + · · · + an ≥ n.
.M
If 1 < k ≤ n + 1, then
a1 a2 an
D
+ + ··· + ≤ 1.
a1k + a2 + · · · + an a1 + a2k + · · · + an a1 + a2 + · · · + ank
A
Solution. Using the substitutions

M
a1 + a2 + · · · + an
s= ,
LY
n
and
O
a1 a an
x1 = , x2 = 2 , · · · , x n = ,
s s s
.M
the desired inequality becomes

x1 xn
+ ··· + ≤ 1,
W
s k−1 x 1k + x2 + · · · + x n x 1 + x 2 + · · · + s k−1 x nk
W
where s ≥ 1 and x 1 + x 2 + · · · + x n = n. Clearly, if this inequality holds for s = 1,

then it holds for any s ≥ 1. Therefore, we need only to consider the case s = 1, when
W
a1 + a2 + · · · + an = n, and the desired inequality is equivalent to

a1 a2 an
+ + ··· + ≤ 1.
a1k − a1 + n a2k − a2 + n ank − an + n
By Bernoulli’s inequality, we have
a1k − a1 + n ≥ 1 + k(a1 − 1) − a1 + n = n − k + 1 + (k − 1)a1 ≥ 0.
Consequently, it suffices to prove that

n
X ai
≤ 1.
i=1
n − k + 1 + (k − 1)ai
For k = n + 1, this inequality is an equality. Otherwise, for 1 < k < n + 1, we rewrite the
inequality as
n
X 1
≥ 1,
i=1
n − k + 1 + (k − 1)ai
which follows from the AM-HM inequality as follows:

n
X 1 n2
≥ Pn = 1.
i=1
n − k + 1 + (k − 1)ai i=1 [n − k + 1 + (k − 1)ai ]
L
D
.M
P 2.107. Let a1 , a2 , . . . , an be nonnegative real numbers such that a1 + a2 + · · · + an ≤ n.
If 0 ≤ k < 1, then
A
a1k
1
+ a2 + · · · + an
+
a1 + a2k
1
+ · · · + an
PI
+ ··· +
1
a1 + a2 + · · · + ank
≥ 1.
M
Solution. By the AM-HM inequality
LY
X 1 n2 n2
≥ =
O
a1k + a2 + · · · + an (a1k + a2 + · · · + an ) a1k + (n − 1) a1

P P P
.M
and Jensen’s inequality

X k
X 1
a1k ≤n a1 ,
W
n
we get
W
X 1 n2
≥ ≥ 1.
a1k + a2 + · · · + an 1
k
+ (n − 1)
P P
W
n n a1 a1
P 2.108. Let a1 , a2 , . . . , an be positive real numbers. If k > 1, then
X ak + ak + · · · + ak n(a1k + a2k + · · · + ank )

2 3 n
≤ .
a2 + a3 + · · · + an a1 + a2 + · · · + an
(Wolfgang Berndt, Vasile Cîrtoaje, 2006)

Solution. Due to homogeneity, we may assume that a1 + a2 + ... + an = 1. Write the

inequality as follows:
X a1

1+ (a2k + a3k + · · · + ank ) ≤ n(a1k + a2k + · · · + ank );
a2 + a3 + · · · + an
X a1 (a k + a k + · · · + a k )
2 3 n
≤ a1k + a2k + · · · + ank ;
a2 + a3 + · · · + an
+ + +
k k k

X a 2 a3 · · · a n
a1 a1k−1 − ≥ 0;
a2 + a3 + · · · + an
L
X a1 a2 (a k−1 − a k−1 ) + a1 a3 (a k−1 − a k−1 ) + · · · + a1 an (a k−1 − a k−1 )
.M
1 2 1 3 1 n
≥ 0;
a2 + a3 + · · · + an
!
aik−1 − a k−1 a k−1 − aik−1
D
j j
X
ai a j + ≥ 0;
1 − ai 1 − aj
A
1≤i< j≤n
X ai a j (aik−1 − a k−1
j
PI
)(ai − a j )
≥ 0.
(1 − ai )(1 − a j )
M
1≤i< j≤n
Since the last inequality is true for k > 1, the proof is finished. The equality holds for
LY
a1 = a2 = · · · = an .
O
.M
W
W
W
L
D
.M
A
PI
M
LY
O
.M
W
W
W
Chapter 3
Symmetric Power-Exponential
L
Inequalities
D
.M
A
3.1 Applications
PI
3.1. If a, b are positive real numbers such that a + b = a4 + b4 , then
M
3 3
aa b b ≤ 1 ≤ aa b b .
LY
O
3.2. If a, b are positive real numbers, then

.M
a2a + b2b ≥ a a+b + b a+b .

W

W
aa + b b ≥ a b + ba .
W
a2a + b2b ≥ a2b + b2a .
3.5. If a, b are nonnegative real numbers such that a + b = 2, then
(a) a b + b a ≤ 1 + a b;
(b) a2b + b2a ≤ 1 + a b.
419
3.6. If a, b are nonnegative real numbers such that a2 + b2 = 2, then
a2b + b2a ≤ 1 + a b.

a+b
a a b b ≤ (a2 − a b + b2 ) 2 .
3.8. If a, b ∈ (0, 1], then
L
a a b b ≤ 1 − a b + a2 b2 .
.M
D
3.9. If a, b are positive real numbers such that a + b ≤ 2, then
A
a b b a
+ ≤ 2. PI
b a
M
LY
3.10. If a, b are positive real numbers such that a + b = 2, then
1
2a a b b ≥ a2b + b2a + (a − b)2 .
O
2
.M
3.11. If a, b ∈ (0, 1] or a, b ∈ [1, ∞), then

W
2a a b b ≥ a2 + b2 .
W
W
2a a b b ≥ a2 + b2 .

a+b
a2 + b2
2
a b
a b ≥ .
2
Symmetric Power-Exponential Inequalities 421
3.14. If a, b are positive real numbers such that a2 + b2 = 2, then

1
2a a b b ≥ a2b + b2a + (a − b)2 .
2
3.15. If a, b ∈ (0, 1], then
1 1

2 2
(a + b ) 2a + 2b ≤ 4.
a b
L
.M
a b b a + 2 ≥ 3a b.
D
A
3.17. If a, b ∈ [0, 1], then
a b−a + b a−b + (a − b)2 ≤ 2.

PI
M
LY
3.18. If a, b are nonnegative real numbers such that a + b ≤ 2, then

O
7
a b−a + b a−b + (a − b)2 ≤ 2.
16
.M
3.19. If a, b are nonnegative real numbers such that a + b ≤ 4, then

W
a b−a + b a−b ≤ 2.
W
W
1
3.20. Let a, b be positive real numbers such that a + b = 2. If k ≥ , then
2
kb ka
a a b b ≥ 1.

p p
a b
a b ≥ 1.
1
1 − a a+1 b b+1 ≥ (1 − a b)2 .
3
a−a + b−b ≤ 2.
L
.M
a2b + b2a ≥ a b + b a ≥ a2 b2 + 1.
D
A
a3b + b3a ≤ 2.
PI
M
LY

O
4
a−b

a3b + b3a + ≤ 2.
2
.M
W

W
2 2
a a + b b ≤ 2.
W

3 3
a a + b b ≥ 2.

2 2
a5b + b5a ≤ 2.

p p
a2 b
+ b2 a
≤ 2.
a b(1 − a b)2 a b(1 − a b)2

≤ a b+1 + b a+1 − 2 ≤ .
2 3
L
.M
a2b + b2a ≤ 1.
D
A
3.33. If a, b are positive real numbers such that a + b = 1, then PI
2a a b b ≥ a2b + b2a .
M
LY

O
a−2a + b−2b ≤ 4.
.M
W
W
W
L
D
.M
A
PI
M
LY
O
.M
W
W
W
3.2 Solutions
P 3.1. If a, b are positive real numbers such that a + b = a4 + b4 , then

3 3
aa b b ≤ 1 ≤ aa b b .
Solution. We will use the inequality
ln x ≤ x − 1, x > 0.
L
.M
To prove this inequality, let us denote
f (x) = x − 1 − ln x, x > 0.
D
From
A
x −1
f 0 (x) = ,
x PI
it follows that f (x) is decreasing on (0, 1] and increasing on [1, ∞). Therefore,
M
f (x) ≥ f (1) = 0.
LY
Using this inequality, we have

O
ln a a b b = a ln a + b ln b ≤ a(a − 1) + b(b − 1) = a2 + b2 − (a + b).

.M
Therefore, the left inequality a a b b ≤ 1 is true if a2 + b2 ≤ a + b. We write this inequality

in the homogeneous form
W
(a2 + b2 )3 ≤ (a + b)2 (a4 + b4 ),

W

W
a b(a − b)(a3 − b3 ) ≥ 0.
1
Taking now x = in the inequality ln x ≤ x − 1 yields
a
a ln a ≥ a − 1.
Similarly,
b ln b ≥ b − 1,
and hence
3 3
ln a a b b = a3 ln a + b3 ln b ≤ a2 (a − 1) + b2 (b − 1) = a3 + b3 − (a2 + b2 ).
3 3
Thus, to prove the right inequality a a b b ≥ 1, it suffices to show that a3 + b3 ≥ a2 + b2 ,
which is equivalent to the homogeneous inequality
(a + b)(a3 + b3 )3 ≥ (a4 + b4 )(a2 + b2 )3 .
We can write this inequality as

A − 3B ≥ 0,
where
A = (a + b)(a9 + b9 ) − (a4 + b4 )(a6 + b6 ),
B = a2 b2 (a2 + b2 )(a4 + b4 ) − a3 b3 (a + b)(a3 + b3 ).
L
Since
.M
A = a b(a3 − b3 )(a5 − b5 ), B = a2 b2 (a − b)(a5 − b5 ),
D
we get
A − 3B = a b(a − b)3 (a5 − b5 ) ≥ 0.
A
Both inequalities become equalities for a = b = 1. PI
M
LY
P 3.2. If a, b are positive real numbers, then

O
a2a + b2b ≥ a a+b + b a+b .

.M
Solution. Assume that a ≥ b and consider the following two cases.

W
Case 1: a ≥ 1. Write the inequality as

W
a a+b (a a−b − 1) ≥ b2b (b a−b − 1).

W
For b ≤ 1, we have
a a+b (a a−b − 1) ≥ 0 ≥ b2b (b a−b − 1).
For b ≥ 1, the inequality is also true since
a a+b ≥ a2b ≥ b2b , a a−b − 1 ≥ b a−b − 1 ≥ 0.
Case 2: a ≤ 1. Since
a2a + b2b ≥ 2a a b b ,
2a a b b ≥ a a+b + b a+b ,
which can be written as a b a

b
+ ≤ 2.
b a
By Bernoulli’s inequality, we get
a
a b a−b b b−a a b(a − b) a(b − a)

b
+ = 1+ + 1+ ≤1+ +1+ = 2.
b a b a b a
The equality holds for a = b.
Conjecture 1. If a, b are positive real numbers, then
L
a4a + b4b ≥ a2a+2b + b2a+2b .
.M
Conjecture 2. If a, b, c are positive real numbers, then
D
a3a + b3b + c 3c ≥ a a+b+c + b a+b+c + c a+b+c .
A
Conjecture 3. If a, b, c, d are positive real numbers, then
PI
a4a + b4b + c 4c + d 4d ≥ a a+b+c+d + b a+b+c+d + c a+b+c+d + d a+b+c+d .
M
LY
O

.M
aa + b b ≥ a b + ba .
(M. Laub, 1985)

W
Solution. Assume that a ≥ b. We will show that if a ≥ 1, then the inequality is true.
From
W
a a−b ≥ b a−b ,
W
we get
a b ba
bb ≥ .
aa
Therefore,
a b ba (a a − a b )(a a − b a )
aa + b b − a b − ba ≥ aa + − a b
− b a
= ≥ 0.
aa aa
Consider further the case 0 < b ≤ a < 1.
First Solution. Denoting
a
c = ab, d = bb, k= ,
b
where c ≥ d and k ≥ 1, the inequality becomes
c k − d k ≥ c − d.
Since the function f (x) = x k is convex for x ≥ 0, from the well-known inequality
f (c) − f (d) ≥ f 0 (d)(c − d),
we get
c k − d k ≥ kd k−1 (c − d).
L
kd k−1 ≥ 1,
.M
b1−a+b ≤ a.
D
Indeed, since 0 < 1 − a + b ≤ 1, by Bernoulli’s inequality, we get
A
b1−a+b = [1 + (b − 1)]1−a+b ≤ 1 + (1 − a + b)(b − 1) = a − b(a − b) ≤ a.
PI
M
Second Solution. Denoting
LY
ba ab a
c= , d= , k= ,
O
a b + ba a b + ba b
where c + d = 1 and k ≥ 1, the inequality becomes
.M
ck a + d k−b ≥ 1.
W
By the weighted AM-GM inequality, we have

W
ck a + d k−b ≥ k ac · k−bd = k ac−bd .

W
Thus, it suffices to show that ac ≥ bd; that is,
a1−b ≥ b1−a ,
which is equivalent to f (a) ≥ f (b), where
ln x
f (x) = .
1− x
It is enough to prove that f (x) is an increasing function. Since
g(x) 1
f 0 (x) = , g(x) = − 1 + ln x.
(1 − x)2 x
we need to show that g(x) ≥ 0 for x ∈ (0, 1). Indeed, from

x −1
g 0 (x) = < 0,
x2
it follows that g(x) is strictly decreasing, hence g(x) > g(1) = 0.
a2a + b2b ≥ a2b + b2a .
L
.M
Solution. Assume that a > b, and denote
a
D
c = ab, d = bb, k= ,
b
A
where c > d and k > 1. The inequality becomes
PI
c 2k − d 2k ≥ c 2 − d 2 .
M
We shall show that
LY
c 2k − d 2k > k(cd)k−1 (c 2 + d 2 ) > c 2 − d 2 .

O
The left inequality follows from Lemma below for x = (c/d)2 . The right inequality is
equivalent to
.M
k(cd)k−1 > 1,
b
(a b)a−b > ,
W
a
1+a− b
W
ln a > ln b.
1−a+ b
W
For fixed a, let us define

1+a− b
f (b) = ln a − ln b, 0 < b < a.
1−a+ b
If f 0 (b) < 0, then f (b) is strictly decreasing, and hence f (b) > f (a) = 0. Since
−2 1
f 0 (b) = ln a − ,
(1 − a + b)2 b
we need to show that g(a) > 0, where
(1 − a + b)2
g(a) = 2 ln a + .
b
From
2 2(1 − a + b) 2(a − b)(a − 1)
g 0 (a) = − = < 0,
a b ab
it follows that g(a) is strictly decreasing on [b, 1), therefore g(a) > g(1) = b > 0. This
completes the proof. The equality holds for a = b.
Lemma. Let k and x be positive real numbers. If either k > 1 and x ≥ 1, or 0 < k < 1
and 0 < x ≤ 1, then
k−1
x k − 1 ≥ k x 2 (x − 1).
Proof. We need to show that f (x) ≥ 0, where
L
k−1
f (x) = x k − 1 − k x 2 (x − 1).
.M
We have
1 k−3 k+1
D
f 0 (x) = k x 2 g(x), g(x) = 2x 2 − (k + 1)x + k − 1.
2
A
Since k−1
g 0 (x) = (k + 1) x 2 − 1 ≥ 0, PI
g(x) is increasing. If x ≥ 1, then g(x) ≥ g(1) = 0, f (x) is increasing, hence f (x) ≥
M
f (1) = 0. If 0 < x ≤ 1, then g(x) ≤ g(1) = 0, f (x) is decreasing, hence f (x) ≥ f (1) =
LY
0. The equality holds for x = 1.
Remark. The following more general inequality holds for 0 < k ≤ e and any positive
O
numbers a and b (Vasile Cîrtoaje, 2006):

.M
a ka + b kb ≥ a kb + b ka .
W
Conjecture 1. If 0 < k ≤ e and a, b ∈ (0, 4], then

W
p
2 a ka b kb ≥ a kb + b ka .
W
Conjecture 2. If a, b ∈ (0, 5], then
2a a b b ≥ a2b + b2a .
Conjecture 3. If a, b ∈ [0, 5], then

a+b
a2 + b2
2
≥ a2b + b2a .
2
Note that Conjecture 1 has been proved for a, b ∈ (0, 1] by A. Coronel and F. Huancas
(2014), and also by L. Matejicka (2014).
P 3.5. If a, b are nonnegative real numbers such that a + b = 2, then
(a) a b + b a ≤ 1 + a b;
(b) a2b + b2a ≤ 1 + a b.

Solution. Without loss of generality, assume that a ≥ b. Since 0 ≤ b ≤ 1 and 0 ≤

a − 1 ≤ 1, by Bernoulli’s inequality, we have
a b ≤ 1 + b(a − 1) = 1 + b − b2
L
and
.M
b a = b · b a−1 ≤ b[1 + (a − 1)(b − 1)] = b2 (2 − b).
(a) We have
D
A
a b + b a − 1 − a b ≤ (1 + b − b2 ) + b2 (2 − b) − 1 − (2 − b)b = −b(b − 1)2 ≤ 0.
PI
The equality holds for a = b = 1, for a = 2 and b = 0, and for a = 0 and b = 2.
M
(b) We have
LY
a2b + b2a − 1 − a b ≤ (1 + b − b2 )2 + b4 (2 − b)2 − 1 − (2 − b)b

= b3 (b − 1)2 (b − 2) = −a b3 (b − 1)2 ≤ 0.
O

.M
W
P 3.6. If a, b are nonnegative real numbers such that a2 + b2 = 2, then

W
a2b + b2a ≤ 1 + a b.
W
Solution. Without loss of generality, assume that a ≥ 1 ≥ b. Applying Bernoulli’s in-

equality gives
a b ≤ 1 + b(a − 1),
hence
a2b ≤ (1 + a b − b)2 .
Also, since 0 ≤ b ≤ 1 and 2a ≥ 2, we have
b2a ≤ b2 .
(1 + a b − b)2 + b2 ≤ 1 + a b,
b(2 + 2a b − a − 2b − a2 b) ≥ 0.
2 + 2a b − a − 2b − a2 b ≥ 0,
L
.M
2(1 − a)(1 − b) + a(1 − a b) ≥ 0,
4(1 − a)(1 − b) + a(a − b)2 ≥ 0.
D
Since a ≥ 1, it suffices to show that
A
4(1 − a)(1 − b) + (a − b)2 ≥ 0. PI
Indeed,
M
4(1 − a)(1 − b) + (a − b)2 = −4(a − 1)(1 − b) + [(a − 1) + (1 − b)]2
LY
= [(a − 1) − (1 − b)]2 = (a + b − 2)2 ≥ 0.

O
p p
.M
W

W
a+b
a a b b ≤ (a2 − a b + b2 ) 2 .
W
Solution. By the weighted AM-GM inequality, we have

a b
a · a + b · b ≥ (a + b)a a+b b a+b ,
a+b
a2 + b2

≥ aa b b .
a+b
2
a2 + b2

2 2
a − ab + b ≥ ,
a+b
(a + b)(a3 + b3 ) ≥ (a2 + b2 )2 ,
a b(a − b)2 ≥ 0.
P 3.8. If a, b ∈ (0, 1], then

a a b b ≤ 1 − a b + a2 b2 .
L
.M
Solution. We claim that
x x ≤ x2 − x + 1
D
for all x ∈ (0, 1]. If this is true, then
A
1 − a b + a2 b2 − a a b b ≥ 1 − a b + a2 b2 − (a2 − a + 1)(b2 − b + 1)
PI
= (a + b)(1 − a)(1 − b) ≥ 0.
To prove the inequality x x ≤ x 2 − x + 1, we show that f (x) ≤ 0 for x ∈ (0, 1], where
M
f (x) = x ln x − ln(x 2 − x + 1).

LY
We have
O
2x − 1
f 0 (x) = ln x + 1 − ,
x2 − x +1
.M
(1 − x)(1 − 2x − x 2 − x 4 )
f 00 (x) = .
x(x 2 − x + 1)2
W
Let x 1 ∈ (0, 1) be the positive root of the equation x 4 + x 2 + 2x = 1. Then, f 00 (x) > 0
for x ∈ (0, x 1 ) and f 00 (x) < 0 for x ∈ (x 1 , 1), hence f 0 is strictly increasing on (0, x 1 ]
W
and strictly decreasing on [x 1 , 1]. Since lim x→0 f 0 (x) = −∞ and f 0 (1) = 0, there is
x 2 ∈ (0, x 1 ) such that f 0 (x 2 ) = 0, f 0 (x) < 0 for x ∈ (0, x 2 ) and f 0 (x) > 0 for x ∈ (x 2 , 1).
W
Therefore, f is decreasing on (0, x 2 ] and increasing on [x 2 , 1]. Since lim x→0 f (x) = 0
and f (1) = 0, it follows that f (x) ≤ 0 for x ∈ (0, 1]. The proof is completed. The
equality holds for a = b = 1.
P 3.9. If a, b are positive real numbers such that a + b ≤ 2, then

a b b a
+ ≤ 2.
b a
Solution. Using the substitution a = t c and b = t d, where c, d, t are positive real

numbers such that c + d = 2 and t ≤ 1, we need to show that
c t d t c
d
+ ≤ 2.
d c
Write this inequality as f (t) ≤ 2, where

c d c
d
t t
f (t) = A + B , A = , B= .
d c
Since f (t) is a convex function, we have
L
.M
f (t) ≤ max{ f (0), f (1)} = max{2, f (1)}.
Therefore, it suffices to show that f (1) ≤ 2; that is,
D
A
2c c d d ≥ c 2 + d 2 .
PI
Setting c = 1 + x and d = 1 − x, where 0 ≤ x < 1, this inequality turns into
M
(1 + x)1+x (1 − x)1−x ≥ 1 + x 2 ,
LY
which is equivalent to f (x) ≥ 0, where

O
f (x) = (1 + x) ln(1 + x) + (1 − x) ln(1 − x) − ln(1 + x 2 ).

.M
We have
2x
f 0 (x) = ln(1 + x) − ln(1 − x) − ,
1 + x2
W
1 1 2(1 − x 2 ) 8x 2
f 00 (x) = + − = .
W
1 + x 1 − x (1 + x 2 )2 (1 − x 2 )(1 + x 2 )2
Since f 00 (x) ≥ 0 for x ∈ [0, 1), it follows that f 0 is increasing, f 0 (x) ≥ f 0 (0) = 0, f (x)
W
is increasing, hence f (x) ≥ f (0) = 0. The proof is completed. The equality holds for
a = b.
P 3.10. If a, b are positive real numbers such that a + b = 2, then
1
2a a b b ≥ a2b + b2a + (a − b)2 .
2
Solution. According to the inequalities in P 3.5-(b) and P 3.9, we have
a2b + b2a ≤ 1 + a b
and
2a a b b ≥ a2 + b2 .
1
a2 + b2 ≥ 1 + a b + (a − b)2 .
2
which is an identity. The equality holds for a = b = 1.
L
.M
D
P 3.11. If a, b ∈ (0, 1] or a, b ∈ [1, ∞), then
A
2a a b b ≥ a2 + b2 .
PI
M
Solution. For a = x and b = 1, the desired inequality becomes
LY
2x x ≥ x 2 + 1, x > 0.
O
If this inequality is true, then

.M
4a a b b − 2(a2 + b2 ) ≥ (a2 + 1)(b2 + 1) − 2(a2 + b2 ) = (a2 − 1)(b2 − 1) ≥ 0.
To prove the inequality 2x x ≥ x 2 + 1, we show that f (x) ≥ 0, where

W
f (x) = ln 2 + x ln x − ln(x 2 + 1).

W
We have
W
2x
f 0 (x) = ln x + 1 − ,
x2 +1
x 2 (x + 1)2 + (x − 1)2
f 00 (x) = .
x(x 2 + 1)2
Since f 00 (x) > 0 for x > 0, f 0 is strictly increasing. Since f 0 (1) = 0, it follows that
f 0 (x) < 0 for x ∈ (0, 1) and f 0 (x) > 0 for x ∈ (1, ∞). Therefore, f is decreasing on
(0, 1] and increasing on [1, ∞), hence f (x) ≥ f (1) = 0 for x > 0. This completes the
proof. The equality holds for a = b = 1.

2a a b b ≥ a2 + b2 .
Solution. By Lemma below, it suffices to show that
(a4 − 2a3 + 4a2 − 2a + 3)(b4 − 2b3 + 4b2 − 2b + 3) ≥ 8(a2 + b2 ),
which is equivalent to A ≥ 0, where
A =a4 b4 − 2a3 b3 (a + b) + 4a2 b2 (a2 + b2 + a b) − [2a b(a3 + b3 ) + 8a2 b2 (a + b)]
+ [3(a4 + b4 ) + 4a b(a2 + b2 ) + 16a2 b2 ] − [6(a3 + b3 ) + 8a b(a + b)]
L
.M
+ 4(a2 + b2 + a b) − 6(a + b) + 9.
We can check that
D
A = [a2 b2 − a b(a + b) + a2 + b2 − 1]2 + B,
A
where PI
B =a2 b2 (a + b)2 − 6a2 b2 (a + b) + [2(a4 + b4 ) + 4a b(a2 + b2 ) + 16a2 b2 ]
M
− [6(a3 + b3 ) + 10a b(a + b)] + [6(a2 + b2 ) + 4a b] − 6(a + b) + 8.
LY
Also, we have
B = [a b(a + b) − 3a b + 1]2 + C,
O
where
.M
C =[2(a4 + b4 ) + 4a b(a2 + b2 ) + 7a2 b2 ] − [6(a3 + b3 ) + 12a b(a + b)]

+ [6(a2 + b2 ) + 10a b] − 6(a + b) + 7,
W
and
C = (a b − 1)2 + 2D,
W
where
W
D =[a4 + b4 + 2a b(a2 + b2 ) + 3a2 b2 ] − [3(a3 + b3 ) + 6a b(a + b)]

+ 3(a + b)2 − 3(a + b) + 3,
It suffices to show that D ≥ 0. Indeed,
D =[(a + b)4 − 2a b(a + b)2 + a2 b2 ] − 3[(a + b)3 − a b(a + b)]
+ 3(a + b)2 − 3(a + b) + 3
=[(a + b)2 − a b]2 − 3(a + b)[(a + b)2 − a b] + 3(a + b)2 − 3(a + b) + 3
2 2
3 a+b

2
= (a + b) − a b − (a + b) + 3 − 1 ≥ 0.
2 2
This completes the proof. The equality holds for a = b = 1.

Lemma. If x > 0, then
1
x x ≥ x + (x − 1)2 (x 2 + 3).
4
Proof. We need to show that f (x) ≥ 0 for x > 0, where
f (x) = ln 4 + x ln x − ln g(x), g(x) = x 4 − 2x 3 + 4x 2 − 2x + 3.
We have
2(2x 3 − 3x 2 + 4x − 1)
f 0 (x) = 1 + ln x − ,
g(x)
L
.M
x 8 + 6x 4 − 32x 3 + 48x 2 − 32x + 9 (x − 1)2 h(x)
f 00 (x) = = ,
g 2 (x) g 2 (x)
D
where
h(x) = x 6 + 2x 5 + 3x 4 + 4x 3 + 11x 2 − 14x + 9.
A
Since PI
h(x) > 7x 2 − 14x + 7 = 7(x − 1)2 ≥ 0,
M
we have f 00 (x) ≥ 0, hence f 0 is strictly increasing on (0, ∞). Since f 0 (1) = 0, it follows
LY
that f 0 (x) < 0 for x ∈ (0, 1) and f 0 (x) > 0 for x ∈ (1, ∞). Therefore, f is decreasing
on (0, 1] and increasing on [1, ∞), hence f (x) ≥ f (1) = 0 for x > 0.
O
.M

W
a+b
a2 + b2
2
aa b b ≥
W
.
2
W
First Solution. Using the substitution a = b x, where x > 0, the inequality becomes as
follows:
bx+b
b x + b2 2
2 2
bx b
(b x) b ≥ ,
2
x+1
b2 x 2 + b2
2
x
(b x) b ≥ ,
2
x+1
x2 + 1
2
b x+1 x x ≥ b x+1 ,
2
x+1
x2 + 1
2
x
x ≥ .
2
It is true if f (x) ≥ 0 for all x > 0, where
x 1 x2 + 1
f (x) = ln x − ln .
x +1 2 2
We have
1 1 x g(x)
f 0 (x) = ln x + − = ,
(x + 1)2 x + 1 x 2 + 1 (x + 1)2
where
L
x2 − 1
g(x) = ln x −
.M
.
x2 + 1
Since
(x 2 − 1)2
D
g 0 (x) = ≥ 0,
x(x 2 + 1)2
A
g is strictly increasing, therefore g(x) < 0 for x ∈ (0, 1), g(1) = 0, g(x) > 0 for x ∈
PI
(1, ∞). Thus, f is decreasing on (0, 1] and increasing on [1, ∞), hence f (x) ≥ f (1) =
0. This completes the proof. The equality holds for a = b.
M
Second Solution. Write the inequality in the form
LY
a+b a2 + b2
a ln a + b ln b ≥ log .
O
2 2
Without loss of generality, consider that a + b = 2k, k > 0, and denote
.M
a = k + x, b = k − x, 0 ≤ x < k.
W
We need to show that f (x) ≥ 0, where

W
f (x) = (k + x) ln(k + x) + (k − x) ln(k − x) − k ln(x 2 + k2 ).

W
We have
2k x
f 0 (x) = ln(k + x) − ln(k − x) − ,
x2+ k2
1 1 2k(x 2 − k2 )
f 00 (x) = + +
k+x k−x (x 2 + k2 )2
2 2
8k x
= 2 .
(k − x 2 )(x 2 + k2 )2
Since f 00 (x) ≥ 0 for x ≥ 0, f 0 is increasing, hence f 0 (x) ≥ f 0 (0) = 0. Therefore, f is

increasing on [0, k), hence f (x) ≥ f (0) = 0.
Remark. For a + b = 2, this inequality can be rewritten as
2a a b b ≥ a2 + b2 ,
a b a
b
2≥ + .
b a
Also, for a + b = 1, the inequality becomes
2a2a b2b ≥ a2 + b2 ,
a 2b 2a
b
2≥ + .
L
b a
.M
D
A
P 3.14. If a, b are positive real numbers such that a2 + b2 = 2, then
1 PI
2a a b b ≥ a2b + b2a + (a − b)2 .
2
M
LY
Solution. According to the inequalities in P 3.6 and P 3.12, we have

O
a2b + b2a ≤ 1 + a b
.M
and
2a a b b ≥ a2 + b2 .
W

W
1
a2 + b2 ≥ 1 + a b + (a − b)2 ,
W
2
which is an identity. The equality holds for a = b = 1.
P 3.15. If a, b ∈ (0, 1], then
1 1

(a2 + b2 ) 2a
+ 2b ≤ 4.
a b

Solution. For a = x and b = 1, the desired inequality becomes
1 + x2
x 2x ≥ , x ∈ (0, 1].
3 − x2
If this inequality is true, it suffices to show that
3 − a2 3 − b2

2 2
(a + b ) + ≤ 4,
1 + a2 1 + b2
L
a2 b2 (2 + a2 + b2 ) + 2 − (a2 + b2 ) − (a2 + b2 )2 ≥ 0,
.M
(2 + a2 + b2 )(1 − a2 )(1 − b2 ) ≥ 0.
D
1 + x2
To prove the inequality x 2x ≥ for 0 < x ≤ 1 , we show that f (x) ≥ 0, where
A
3 − x2
f (x) = x ln x +
1
2
1PI
ln(3 − x 2 ) − ln(1 + x 2 ).
2
M
We have
x x
LY
f 0 (x) = 1 + ln x − − ,
3− x 2 1 + x2
O
1 3 + x2 1 − x2
f 00 (x) = − −
.M
x (3 − x 2 )2 (1 + x 2 )2
(1 − x)(9 + 6x − x 3 ) 1 − x2
= − .
x(3 − x)2 (1 + x 2 )2
W
We will show that f 00 (x) > 0 for 0 < x < 1. This is true if
W
9 + 6x − x 3 1+ x
W
− > 0.
x(3 − x)2 (1 + x 2 )2
Indeed,
9 + 6x − x 3 1+ x 9 1+ x 1
− > − = > 0.
x(3 − x)2 (1 + x )
2 2 9x x(1 + x)2 1+ x
Since f 00 (x) > 0, f 0 is strictly increasing on (0, 1]. Since f 0 (1) = 0, it follows that
f 0 (x) < 0 for x ∈ (0, 1), f is strictly decreasing on (0, 1], hence f (x) ≥ f (1) = 0. This
completes the proof. The equality holds for a = b = 1.
a b b a + 2 ≥ 3a b.
Solution. Setting a = 1+ x and b = 1− x, where 0 ≤ x < 1, the inequality is equivalent

to
(1 + x)1−x (1 − x)1+x ≥ 1 − 3x 2 .
1
Consider further the non-trivial case 0 ≤ x < p , and write the desired inequality as
3
f (x) ≥ 0, where
L
.M
f (x) = (1 − x) ln(1 + x) + (1 + x) ln(1 − x) − ln(1 − 3x 2 ).
We have
D
1− x 1+ x 6x
f 0 (x) = − ln(1 + x) + ln(1 − x) + − + ,
A
1 + x 1 − x 1 − 3x 2
1 00 −1 2(x 2 + 1) 3(3x 2 + 1)
PI
f (x) = − + .
2 1 − x 2 (1 − x 2 )2 (1 − 3x 2 )2
M
1
Making the substitution t = x 2 , 0 ≤ t < , we get
3
LY
1 00 3(3t + 1) t +3 4t(5 − 9t)

f (x) = − = > 0.
O
2 (3t − 1)2 (t − 1)2 (t − 1)2 (3t − 1)2
Therefore, f 0 (x) is strictly increasing, f 0 (x) ≥ f 0 (0) = 0, f (x) is strictly increasing, and
.M
hence f (x) ≥ f (0) = 0. This completes the proof. The equality holds for a = b = 1.
W
W
P 3.17. If a, b ∈ [0, 1], then

W
a b−a + b a−b + (a − b)2 ≤ 2.
Solution (by Vo Quoc Ba Can). Without loss of generality, assume that a ≥ b. Using the
substitution
c = a − b,
(b + c)−c + b c + c 2 ≤ 2
for
0 ≤ b ≤ 1 − c, 0 ≤ c ≤ 1.
If c = 1, then b = 0, and the inequality is an equality. Also, for c = 0, the inequality is

an equality. Consider further that 0 < c < 1. We need to show that f (x) ≤ 0, where
f (x) = (x + c)−c + x c + c 2 − 2, x ∈ [0, 1 − c].
We claim that f 0 (x) > 0 for x > 0. Then, f (x) is strictly increasing on [0, 1 − c], hence
f (x) ≤ f (1 − c) = (1 − c)c − (1 − c 2 ).
In addition, by Bernoulli’s inequality,
f (x) ≤ 1 + c(−c) − (1 − c 2 ) = 0.
L
.M
Since
c[(x + c)1+c − x 1−c ]
f 0 (x) = ,
(x + c)1+c x 1−c
D
we have f 0 (x) > 0 for x > 0 if and only if
A
1−c
x + c > x 1+c . PI
M
2c
Let d = 1−c . Using the weighted AM-GM inequality, we have
LY
1 1+c
1 − c d 1+d 1 + c 1 − c 1−c 1−c

1−c

x +c =1· x +d · ≥ (1 + d) x = x 1+c .
2 2 2 2
O

.M
1−c
1+c 1 − c 1+c

≥ .
2 2
W
Indeed, using Bernoulli’s inequality, we get

W
1−c 1−c
1−c 1 + c 1+c 1−c 1+c 1+c
1+c
W
= 1− ≤1− · = .
2 2 1+c 2 2
The equality holds for a = b, for a = 1 and b = 0, and for a = 0 and b = 1.
P 3.18. If a, b are nonnegative real numbers such that a + b ≤ 2, then

7
a b−a + b a−b + (a − b)2 ≤ 2.
16
Solution. Assume that a > b and use the substitution
c = a − b.

7 2
a−c + (a − c)c + c ≤2
16
for
c
0 < c ≤ 2, c ≤ a ≤1+ .
2
Equivalently, we need to show that f (x) ≤ 0 for
ci
L
h
x ∈ c, 1 + , 0<c≤2,
2
.M
where
7 2
f (x) = x −c + (x − c)c +
D
c − 2.
16
A
As we have shown
h in the preceding proof, f 0 (x) > 0 for x > c. Therefore, f (x) is strictly
c i
increasing on c, 1 + , and hence
2
PI
M
c c −c c c 7 2
f (x) ≤ f 1 + = 1+ + 1− + c − 2.
2 2 2 16
LY
To end the proof we need to show that

O
c −c c c 7 2
1+ + 1− + c ≤2
2 2 16
.M
for c ∈ [0, 2]. Indeed, by Lemma 1 and Lemma 2 below, we have

c −c
W
3 2
1+ + c ≤ 1,
2 16
W
c c 1 2
1− + c ≤ 1.
2 4
W
Adding these inequalities yields the desired inequality. The equality holds for a = b, for
a = 2 and b = 0, and for a = 0 and b = 2.
Lemma 1. If 0 ≤ x ≤ 2, then
x −x 3 2
1+ + x ≤ 1,
2 16
with equality for x = 0 and x = 2.
Proof. We need to show that f (x) ≤ 0 for 0 ≤ x ≤ 2, where
3 2

x
f (x) = −x ln 1 + − ln 1 − x .
2 16
We have
x x(3x 2 + 6x − 4)
f 0 (x) = − ln 1 + + ,
2 (2 + x)(16 − 3x 2 )
g(x)
f 00 (x) = ,
(2 + x)2 (16 − 3x 2 )2
where
g(x) = −9x 5 − 18x 4 + 168x 3 + 552x 2 + 128x − 640.
Since g(x 1 ) = 0 for x 1 ≈ 0, 88067, g(x) < 0 for x ∈ [0, x 1 ) and g(x) > 0 for x ∈ (x 1 , 2],
f 0 is strictly decreasing on [0, x 1 ] and strictly increasing on [x 1 , 2]. Since f 0 (0) = 0
5
and f 0 (2) = − ln 2 + > 0, there is x 2 ∈ (x 1 , 2) such that f 0 (x 2 ) = 0, f 0 (x) < 0 for
L
2
x ∈ (0, x 2 ), and f 0 (x) > 0 for x ∈ (x 2 , 2]. Therefore, f is decreasing on [0, x 2 ] and
.M
increasing on [x 2 , 2]. Since f (0) = f (2) = 0, it follows that f (x) ≤ 0 for x ∈ [0, 2].
D
Lemma 2. If 0 ≤ x ≤ 2, then
A
x x 1 2
1− + x ≤ 1,
with equality for x = 0 and x = 2.

2 4 PI
M
Proof. We need to show that f (x) ≤ 0, where
LY
1 2

x
f (x) = x ln 1 − − ln 1 − x .
2 4
O
We have
x x2
.M
f 0 (x) = ln 1 − − ,
2 4 − x2
−1 8x
f 00 (x) = − .
W
2 − x (4 − x 2 )2
Since f 00 < 0 for 0 ≤ x < 2, f 0 is strictly decreasing, hence f 0 (x) ≤ f 0 (0) = 0, f is
W
strictly decreasing, therefore f (x) ≤ f (0) = 0 for x ∈ [0, 2].

W
1
Conjecture. If a, b are nonnegative real numbers such that a + b = , then
4
a2(b−a) + b2(a−b) ≤ 2.
P 3.19. If a, b are nonnegative real numbers such that a + b ≤ 4, then
a b−a + b a−b ≤ 2.

Solution. Without loss of generality, assume that a ≥ b. Consider first that a − b ≥ 2.

We have
a ≥ a − b ≥ 2,
and from
4 ≥ a + b = (a − b) + 2b ≥ 2 + 2b,
we get b ≤ 1. Clearly, the desired inequality is true because
a b−a < 1, b a−b ≤ 1.
Since the case a − b = 0 is trivial, consider further that 0 < a − b < 2 and use the
L
substitution
c = a − b.
.M
a−c + (a − c)c ≤ 2
D
A
for
c
0 < c < 2, c ≤ a ≤2+ . PI
2
Equivalently, we need to show that f (x) ≤ 0 for
M
h ci
x ∈ c, 2 + , 0<c<2,
LY
2
where
O
f (x) = x −c + (x − c)c − 2.
.M
The derivative
c[x 1+c − (x − c)1−c ]
f 0 (x) =
x 1+c (x − c)1−c
W
has the same sign as

W
g(x) = (1 + c) ln x − (1 − c) ln(x − c).

W
We have
c(2x − 1 − c)
g 0 (x) = .
x(x − c)
5

Case 1: c = 1. We need to show that f (x) ≤ 0 for x ∈ 1, , where
2
x 2 − 3x + 1
f (x) = .
x
Indeed, we have
(x − 1)(2x − 5) + (x − 3)
f (x) = < 0.
2x
Case 2: 1 < c < 2. Since

2x − 1 − c ≥ 2c − 1 − c > 0,
we have g 0 (x) > 0, hence g(x) is strictly increasing. Since g(x) → −∞ when x → c
and c c c
g 2+ = (1 + c) ln 2 + + (c − 1) ln 2 −
2 2 2
c2

c c
> (c − 1) ln 2 + + (c − 1) ln 2 − = (c − 1) ln 4 − > 0,
2 2 4
c
L
there exists x 1 ∈ c, 2 + such that g(x 1 ) = 0, g(x) < 0 for x ∈ (c, x 1 ) and g(x) > 0
2
.M
c h ci
for x ∈ x 1 , 2 + . Thus, f (x) is decreasing on [c, x 1 ] and increasing on x 1 , 2 + .
2 2
c
Then, it suffices to show that f (c) ≤ 0 and f 2 +
D
≤ 0. We have
2
A
f (c) = c −c − 2 < 1 − 2 < 0.

Also, the inequality f 2 +
c
≤ 0 has the form
PI
M
2
c −c c c
LY

2+ + 2− ≤ 2,
2 2
O
which follows immediately from Lemma 1 below.

h ci
.M
Case 3: 0 < c < 1. We claim that g(x) > 0 for x ∈ c, 2 + . From

2
c(2x − 1 − c)
g 0 (x) =
W
,
x(x − c)
W
1+c 1+c

c
it follows that g(x) is decreasing on c, , and increasing on , 2 + . Then,
2 2 2
W
1+c 1+c 1−c

g(x) ≥ g = (1 + c) ln − (1 − c) ln ,
2 2 2
1+c 1−c
1+c 1−c

> .
2 2
This inequality follows from Bernoulli’s inequality, as follows
1+c
1+c 1 − c 1+c (1 + c)(1 − c) 1 + c 2

= 1− >1− =
2 2 2 2
and 1−c
1−c 1 + c 1−c (1 − c)(1 + c) 1 + c 2

= 1− <1− = .
2 2 2 2
h ci
Since g(x) > 0 involves f 0 (x) > 0, it follows that f (x) is strictly increasing on c, 2 + ,
2
and hence c
f (x) ≤ f 2 + .
2
c
So, we need to show that f 2 + ≤ 0 for 0 < c < 2, which follows immediately from
2
Lemma 1.
The proof is completed. The equality holds for a = b.
L
.M
Lemma 1. If 0 ≤ c ≤ 2, then
c −c c c
2+ + 2−
D
≤ 2.
2 2
A
Proof. According to Lemma 2, the following inequalities hold for c ∈ [0, 2]:
c −c
PI
c3
2+ ≤ 1 − c ln 2 + ,
M
2 9
LY
c c c3
2− ≤ 1 + c ln 2 − .
2 9
O
Summing these inequalities, the desired inequality follows.

.M
Lemma 2. If
−2 ≤ x ≤ 2,
W
then
x x x3
2− ≤ 1 + x ln 2 − .
W
2 9
Proof. We have
W
ln 2 ≈ 0.693 < 7/9.

If x ∈ [0, 2], then
x3 x3 8
1 + x ln 2 − ≥1− ≥ 1 − > 0.
9 9 9
Also, if x ∈ [−2, 0], then
x3 7x x 3 8 + 7x − x 3
1 + x ln 2 − ≥1+ − >
9 9 9 9
2(x + 2)2 + (−x)(x + 1)2
= > 0.
9
So, we can write the desired inequality as f (x) ≥ 0, where
x3
x
f (x) = ln 1 + d x − − x ln 2 − , d = ln 2.
9 2
We have
9d − 3x 2 x x
f 0 (x) = + − ln 2 − .
9 + 9d x − x 3 4 − x 2
Since f (0) = 0, it suffices to show that f 0 (x) ≤ 0 for x ∈ [−2, 0], and f 0 (x) ≥ 0 for
x ∈ [0, 2]; that is, x f 0 (x) ≤ 0 for x ∈ [−2, 2]. According to Lemma 3, the inequality
x f 0 (x) ≤ 0 is true if x g(x) ≥ 0, where
L
9d − 3x 2 1 2

x x
.M
g(x) = + − d− − x .
9 + 9d x − x 3 4 − x 4 32
D
We have
A
9d − 3x 2

1 2

x x
g(x) = −d + + + x
9 + 9d x − x 3
2
4 − x 4 32
d x − 3x − 9d 2 64 − 4x − x 2
PI

=x +
M
,
9 + 9d x − x 3 32(4 − x)
LY
hence
O
x 2 g1 (x)
x g(x) = ,
32(4 − x)(9 + 9d x − x 3 )
.M
where
W
g1 (x) =32(4 − x)(d x 2 − 3x − 9d 2 ) + (64 − 4x − x 2 )(9 + 9d x − x 3 )

=x 5 + 4x 4 − (64 + 41d)x 3 + (87 + 92d)x 2 + 12(24d 2 + 48d − 35)x
W
+ 576(1 − 2d 2 ).
W
Since g1 (x) ≥ 0 for x ∈ [a1 , b1 ], where a1 ≈ −12.384 and b1 =≈ 2.652, we have

g1 (x) ≥ 0 for x ∈ [−2, 2].
Lemma 3. If x < 4, then
xh(x) ≤ 0,
where
1 2

x x
h(x) = ln 2 − − ln 2 − − x .
2 4 32
Proof. From
−x 2
h0 (x) = ≤ 0,
16(4 − x)
it follows that h(x) is decreasing. Since h(0) = 0, we have h(x) ≥ 0 for x ≤ 0, and
h(x) ≤ 0 for x ∈ [0, 4); that is, xh(x) ≤ 0 for x < 4.
1
P 3.20. Let a, b be positive real numbers such that a + b = 2. If k ≥ , then
2
kb ka
a a b b ≥ 1.
L
.M
Solution. Setting a = 1+ x and b = 1− x, where 0 ≤ x < 1, the inequality is equivalent
to
(1 + x)k(1−x) ln(1 + x) + (1 − x)k(1+x) ln(1 − x) ≥ 0.
D
Consider further the non-trivial case 0 < x < 1, and write the desired inequality as
A
f (x) ≥ 0, where
PI
f (x) = k(1 − x) ln(1 + x) − k(1 + x) ln(1 − x) + ln ln(1 + x) − ln(− ln(1 − x)).
M
We have
LY
2k(1 + x 2 ) 1 1
f 0 (x) = − k ln(1 − x 2 ) + +
1 − x2 (1 + x) ln(1 + x) (1 − x) ln(1 − x)
O
2k 1 1
> + +
.M
1− x 2 (1 + x) ln(1 + x) (1 − x) ln(1 − x)
1 1 1
≥ + +
1− x 2 (1 + x) ln(1 + x) (1 − x) ln(1 − x)
W
g(x)
= ,
(1 − x ) ln(1 + x) ln(1 − x)
W
where
W
g(x) = ln(1 + x) ln(1 − x) + (1 + x) ln(1 + x) + (1 − x) ln(1 − x).

It suffices to show that f 0 (x) > 0. Indeed, if this is true, then f (x) is strictly increasing,
hence
f (x) > lim f (x) = 0.
x→0
Thus, we need to prove that g(x) < 0. We have

−x
g 0 (x) = h(x),
1 − x2
where
h(x) = (1 + x) ln(1 + x) + (1 − x) ln(1 − x).
Since
1+ x
h0 (x) = ln > 0,
1− x
h(x) is strictly increasing, h(x) > h(0) = 0, g 0 (x) < 0, g(x) is strictly decreasing, and
hence g(x) < g(0) = 0. This completes the proof. The equality holds for a = b = 1.

p p
a b
a b ≥ 1.
L
.M
Solution. For a = b = 1, the equality holds. In what follows, we will assume that
D
a > 1 > b. Taking logarithms of both sides, the inequality becomes in succession
A
p p
a ln a + b ln b ≥ 0,
p p
PI
a ln a ≥ b(− ln b),
M
1 1
ln a + ln ln a ≥ ln b + ln(− ln b).
LY
2 2
Setting a = 1 + x and b = 1 − x, we need to show that f (x) ≥ 0 for 0 < x < 1, where
O
1 1
f (x) = ln(1 + x) − ln(1 − x) + ln ln(1 + x) − ln(− ln(1 − x)).
.M
2 2
We have
1 1 1
W
f 0 (x) = + + .
1− x 2 (1 + x) ln(1 + x) (1 − x) ln(1 − x)
W
As shown in the proof of the preceding P 3.20, we have f 0 (x) > 0. Therefore, f (x) is
strictly increasing and
W
f (x) > lim f (x) = 0.

x→0
1
1 − a a+1 b b+1 ≥ (1 − a b)2 .
3
Solution. Putting a = 1 + x and b = 1 − x, 0 ≤ x < 1, the inequality becomes

1 4
(1 + x)2+x (1 − x)2−x ≤ 1 − x .
3
Write this inequality as f (x) ≤ 0, where
1

f (x) = (2 + x) ln(1 + x) + (2 − x) ln(1 − x) − ln 1 − x 4 .
3
We have
2x 4x 3
f 0 (x) = ln(1 + x) − ln(1 − x) − + ,
1 − x2 3 − x4
L
2 2(1 + x 2 ) 4x 2 (x 4 + 9)
.M
f 00 (x) = − +
1 − x 2 (1 − x 2 )2 (3 − x 4 )2
−4x 2 4x 2 (x 4 + 9) −8x 4 [x 4 + 1 + 8(1 − x 2 )]
D
= + = ≤ 0.
(1 − x 2 )2 (3 − x 4 )2 (1 − x 2 )2 (3 − x 4 )2
A
Therefore, f 0 (x) is decreasing, f 0 (x) ≤ f 0 (0) = 0, f (x) is decreasing, f (x) ≤ f (0) = 0.
The equality holds for a = b = 1. PI
M
LY

O
a−a + b−b ≤ 2.
.M
Solution. Consider that a ≥ b, when

W
0 < b ≤ 1 ≤ a < 2,
W
and write the inequality as

W
aa − 1 b b − 1
+ ≥ 0.
aa bb
According to Lemma from the proof of P 3.4, we have
a+1 b+1
aa − 1 ≥ a 2 (a − 1), bb − 1 ≥ b 2 (b − 1).

1−a 1−b
a 2 (a − 1) + b 2 (b − 1) ≥ 0,
1−a 1−b
a 2 ≥b 2 ,
1−b
(a b) 2 ≤ 1,
a b ≤ 1.
The last inequality follows immediately from the AM-GM inequality
1
ab ≤ (a + b)2 = 1.
4
L
.M
a2b + b2a ≥ a b + b a ≥ a2 b2 + 1.
D
A
Solution. Since a, b ∈ [0, 2] and (1 − a)(1 − b) ≤ 0, from Lemma below, we have
PI
b(a b + 3 − a − b)(a − 1) b(a b + 1)(a − 1)
ab − 1 ≥ =
M
2 2
LY
and
a(a b + 1)(b − 1)
ba − 1 ≥ .
O
2
Based on these inequalities, we get
.M
a b + b a − a2 b2 − 1 = (a b − 1) + (b a − 1) + 1 − a2 b2
b(a b + 1)(a − 1) a(a b + 1)(b − 1)
W
≥ + + 1 − a2 b2
2 2
= (ab + 1)(a b − 1) + 1 − a2 b2 = 0
W
and
W
a2b + b2a − a b − b a = a b (a b − 1) + b a (b a − 1)
a b b(a b + 1)(a − 1) b a a(a b + 1)(b − 1)
≥ +
2 2
a b(a b + 1)(a − b)(a b−1 − b a−1 )
= .
4
On the valid assumption a ≥ b, we only need to show that a b−1 ≥ b a−1 , which is
equivalent to
b−a a−b a−b
a 2 ≥b 2 , 1 ≥ (a b) 2 , 1 ≥ a b, (a − b)2 ≥ 0.
For both inequalities, the equality holds when a = b = 1, when a = 0 and b = 2, and
when a = 2 and b = 0.
Lemma. If x, y ∈ [0, 2] such that (1 − x)(1 − y) ≤ 0, then
y(x y + 3 − x − y)(x − 1)
xy −1≥ ,
2
with equality for x = 1, for y = 0, for y = 1, and for x = 0 and y = 2.
Proof. For y = 0, y = 1 and y = 2, the inequality is an identity. For fixed y ∈ (0, 1) ∪
(1, 2), let us define
L
y(x y + 3 − x − y)(x − 1)
.M
f (x) = x y − 1 − .
2
D
We have
x y + 3 − x − y (x − 1)( y − 1)

0
f (x) = y x y−1
− − ,
A
2 2
PI
f 00 (x) = y( y − 1)(x y−2 − 1).
M
Since f 00 (x) ≥ 0 for x ∈ (0, 2], f 0 is increasing. There are two cases to consider.
LY
Case 1: x ≥ 1 > y. We have f 0 (x) ≥ f 0 (1) = 0, f (x) is increasing, hence f (x) ≥ f (1) =
0.
O
Case 2: y > 1 ≥ x. We have f 0 (x) ≤ f 0 (1) = 0, f (x) is decreasing, hence f (x) ≥

f (1) = 0.
.M
W

W
a3b + b3a ≤ 2.
W
Solution. Without loss of generality, assume that a ≥ b. Using the substitution a = 1+ x

and b = 1 − x, 0 ≤ x < 1, we can write the inequality as
e3(1−x) ln(1+x) + e3(1+x) ln(1−x) ≤ 2.
Applying Lemma below, it suffices to show that f (x) ≤ 2, where
x2 x3 x2 x3
f (x) = e3(1−x)(x− 2 + 3 ) + e−3(1+x)(x+ 2 + 3 ) .
Since f (0) = 2, it suffices to show that f 0 (x) ≤ 0 for x ∈ [0, 1). From
15 2 9x 2 5x 3 4
f 0 (x) =(3 − 9x + x − 4x 3 )e3x− 2 + 2 −x
2
15 2 9x 2 5x 3 4
−(3 + 9x + x + 4x 3 )e−3x− 2 − 2 −x ,
2
it follows that f 0 (x) ≤ 0 is equivalent to
3 6 − 18x + 15x 2 − 8x 3
e−6x−5x ≥ .
6 + 18x + 15x 2 + 8x 3
For the nontrivial case 6 − 18x + 15x 2 − 8x 3 > 0, we rewrite this inequality as g(x) ≥ 0,
L
where
.M
g(x) = −6x − 5x 3 − ln(6 − 18x + 15x 2 − 8x 3 ) + ln(6 + 18x + 15x 2 + 8x 3 ).
D
Since g(0) = 0, it suffices to show that g 0 (x) ≥ 0 for x ∈ [0, 1). From
A
1 0 (6 + 8x 2 ) − 10x (6 + 8x 2 ) + 10x
g (x) = −2 − 5x 2 + + ,
3 PI
6 + 15x 2 − (18x + 8x 3 ) 6 + 15x 2 + (18x + 8x 3 )
it follows that g 0 (x) ≥ 0 is equivalent to
M
2(6 + 8x 2 )(6 + 15x 2 ) − 20x(18x + 8x 3 ) ≥ (2 + 5x 2 )[(6 + 15x 2 )2 − (18x + 8x 3 )2 ].
LY
Since
O
(6 + 15x 2 )2 − (18x + 8x 3 )2 ≤ (6 + 15x 2 )2 − 324x 2 − 288x 4 ≤ 4(9 − 36x 2 ),

.M
(3 + 4x 2 )(6 + 15x 2 ) − 5x(18x + 8x 3 ) ≥ (2 + 5x 2 )(9 − 36x 2 ).

W
This reduces to 6x 2 +200x 4 ≥ 0, which is clearly true. The equality holds for a = b = 1.
W
Lemma. If t > −1, then

W
t2 t3
ln(1 + t) ≤ t − + .
2 3
Proof. We need to prove that f (t) ≥ 0, where
t2 t3
f (t) = t − + − ln(1 + t).
2 3
Since
t3
f 0 (t) =,
t +1
f (t) is decreasing on (−1, 0] and increasing on [0, ∞). Therefore, f (t) ≥ f (0) = 0.

4
a−b

a3b + b3a + ≤ 2.
2
Solution (by M. Miyagi and Y. Nishizawa). We may assume that a = 1+ x and b = 1− x,

where 0 ≤ x ≤ 1. The desired inequality is equivalent to
(1 + x)3(1−x) + (1 − x)3(1+x) + x 4 ≤ 2.
L
By Lemma below, we have
.M
1
(1 + x 1−x ≤ (1 + x)2 (2 − x 2 )(2 − 2x + x 2 ),
4
D
1
A
(1 − x 1+x ≤ (1 − x)2 (2 − x 2 )(2 + 2x + x 2 ).
4
PI
M
(1 + x)6 (2 − x 2 )3 (2 − 2x + x 2 )3 + (1 − x)6 (2 − x 2 )3 (2 + 2x + x 2 )3 + 64x 2 ≤ 128,
LY
O
x 4 (1 − x 2 )[x 6 (x 6 − 8x 4 + 31x 2 − 34) − 2(17x 6 − 38x 4 + 16x 2 + 8)] ≤ 0.

.M
t 3 − 8t 2 + 31t − 34 < 0
W
and
W
17t 3 − 38t 2 + 16t + 8 > 0

W
for all t ∈ [0, 1]. Indeed, we have
t 3 − 8t 2 + 31t − 34 < t 3 − 8t 2 + 31t − 24 = (t − 1)(t 2 − 7t + 24) ≤ 0,
17t 3 − 38t 2 + 16t + 8 = 17t(t − 1)2 + (−4t 2 − t + 8) > 0.
Lemma. If −1 ≤ t ≤ 1, then
1
(1 + t)1−t ≤ (1 + t)2 (2 − t 2 )(2 − 2t + t 2 ),
4
with equality for t = −1, t = 0 and t = 1.
Proof. We need to prove that f (t) ≥ 0 for −1 < t ≤ 1, where
f (t) = (1 + t) ln(1 + t) + ln(2 − t 2 ) + ln(2 − 2t + t 2 ) − ln 4.
We have
2t 2(t − 1)
f 0 (t) = 1 + ln(1 + t) − + ,
2 − t 2 2 − 2t + t 2
t 2 g(t)
f 00 (t) = ,
(1 + t)(2 − t 2 )2 (2 − 2t + t 2 )2
where
g(t) = t 6 − 8t 5 + 12t 4 + 8t 3 − 20t 2 − 16t + 16.
L
Case 1: 0 ≤ t ≤ 1. From
.M
g 0 (t) = 6t 5 − 40t 4 + 48t 3 + 24t 2 − 40t − 16 = 6t 5 − 8t − 16 − 8t(5t 3 − 6t 2 − 3t + 4)
D
= (6t 5 − 8t − 16) − 8t(t − 1)2 (5t + 4) < 0,
A
it follows that g is strictly decreasing on [0, 1]. Since g(0) = 16 and g(1) = −7, there
PI
exists a number c ∈ (0, 1) such that g(c) = 0, g(t) > 0 for 0 ≤ t < c and g(t) < 0
for c < t ≤ 1. Therefore, f 0 is strictly increasing on [0, c] and strictly decreasing on
M
[c, 1]. From f 0 (0) = 0 and f 0 (1) = ln 2 − 1 < 0, it follows that there exists a number
d ∈ (0, 1) such that f 0 (d) = 0, f 0 (t) > 0 for 0 < t < d and f 0 (t) < 0 for c < t ≤ 1. As
LY
a consequence, f is strictly increasing on [0, d] and strictly decreasing on [d, 1]. Since
f (0) = 0 and f (1) = 0, we have f (t) ≥ 0 for 0 ≤ t ≤ 1.
O
Case 2: −1 < t ≤ 0. From

.M
g(t) = t 4 (t − 2)(t − 6) + 4(t + 1)(2t 2 − 7t + 3) + 4 > 0,

W
it follows that f 0 is strictly increasing on (−1, 0]. Since f 0 (0) = 0, we have f 0 (t) < 0
for −1 < t < 0, hence f is strictly decreasing on (−1, 0]. From f (0) = 0, it follows that
W
f (t) ≥ 0 for −1 < t ≤ 0.

Conjecture. If a, b are nonnegative real numbers such that a + b = 2, then
W
a−b 2

a3b + b3a + ≥ 2.
2

2 2
a a + b b ≤ 2.

Solution. Without loss of generality, assume that
0 < a ≤ 1 ≤ b < 2,
and write the inequality as

1 1
+ ≤ 2.
1 1/a 1 2/b

a 2 b
By Bernoulli’s inequality, we have
1 1/a 1 1 a3 − a2 + 1

≥ 1 + − 1 = ,
a2 a a2 a3
L
2/b
b2 − 2b + 2
.M
1 2 1

≥1+ −1 = .
b b b b2
D
A
a3 b2
+ ≤ 2,
a3 − a2 + 1 b2 − 2b + 2 PI
M
a3 (2 − b)2
≤ ,
a3 − a2 + 1 b2 − 2b + 2
LY
a3 a2
≤ ,
a3 − a2 + 1 a2 − 2a + 2
O
a2 (a − 1)2 ≥ 0.
.M
The equality happens for a = b = 1.

W

W
3 3
a a + b b ≥ 2.
W
Solution. Assume that a ≤ b; that is,
0 < a ≤ 1 ≤ b < 2.
3 3
There are two cases to consider: 0 < a ≤ and ≤ a ≤ 1.
5 5
3 7
Case 1: 0 < a ≤ . From a + b = 2, we get ≤ b < 2. Let
5 5
3
f (x) = x x , 0 < x < 2.
Since
3
f 0 (x) = 3x x −2 (1 − ln x) > 0,
7 7

f (x) is increasing on , 2 , and hence f (b) ≥ f ; that is,
5 5
15/7
3 7
bb ≥ .
5
Using Bernoulli’s inequality gives

15/7
7 7 2 8/7 7 16 51

3
= 1+ > 1+ = > 2.
L
b ≥
b
5 5 5 5 35 25
.M
hence
3 3
a a + b b > 2.
D
A
3
Case 2:
5
≤ a ≤ 1. By Lemma below, we have PI
M
3
2a a ≥ 3 − 15a + 21a2 − 7a3
LY
and
3
2b b ≥ 3 − 15b + 21b2 − 7b3 .
O
Summing these inequalities, we get

.M
3 3

2 a a + b b ≥ 6 − 15(a + b) + 21(a2 + b2 ) − 7(a3 + b3 )
W
= 6 − 15(a + b) + 21(a + b)2 − 7(a + b)3 = 4.

W
This completes the proof. The equality holds for a = b = 1.

3
W
Lemma. If ≤ x ≤ 2, then
5
3
2x x ≥ 3 − 15x + 21x 2 − 7x 3 ,
with equality for x = 1.

Proof. We first show that h(x) > 0, where
h(x) = 3 − 15x + 21x 2 − 7x 3 .
From
h0 (x) = 3(−5 + 14x − 7x 2 ),
s s s
2 2 2
it follows that h(x) is increasing on 1 − ,1 + , and decreasing on 1 + ,∞ .
7 7 7
3

Then, it suffices to show that f ≥ 0 and f (2) ≥ 0. Indeed
5
3 6

f = , f (2) = 1.
5 125
Write now the desired inequality as f (x) ≥ 0, where
3 3
f (x) = ln 2 + ln x − ln(3 − 15x + 21x 2 − 7x 3 ), ≤ x ≤ 2.
x 5
L
.M
We have
x2 0 x 2 (7x 2 − 14x + 5)
f (x) = g(x), g(x) = 1 − ln x + ,
3 3 − 15x + 21x 2 − 7x 3
D
g1 (x)
A
g 0 (x) = ,
(3 − 15x + 21x 2 − 7x 3 )2
where
PI
M
g1 (x) = −49x 7 + 245x 6 − 280x 5 − 147x 4 + 471x 3 − 321x 2 + 90x − 9.
LY
In addition,
O
g1 (x = (x − 1)2 g2 (x), g2 (x) = −49x 5 + 147x 4 + 63x 3 − 168x 2 + 72x − 9,

.M
g2 (x) = 11x 5 + 3g3 (x), g3 (x) = −20x 5 + 49x 4 + 21x 3 − 56x 2 + 24x − 3,
W
g3 (x) = (4x − 1)g4 (x), g4 (x) = −5x 4 + 11x 3 + 8x 2 − 12x + 3,

W
g4 (x) = x 5 + g5 (x), g5 (x) = −6x 4 + 11x 3 + 8x 2 − 12x + 3,
g5 (x) = (2x − 1)g6 (x), g6 (x) = −3x 3 + 4x 2 + 6x − 3,

W
g6 (x) = 1 + (2 − x)(3x 2 + 2x − 2).
Therefore, we get in succession g6 (x) > 0, g5 (x) > 0, g4 (x) > 0, g3 (x) > 0, g2 (x)
> 0,
3

0
g1 (x) ≥ 0, g (x) ≥ 0, g(x) is increasing. Since g(1) = 0, we have g(x) < 0 on ,1
5
3

and g(x) > 0 on (1, 2]. Then, f (x) is decreasing on , 1 and increasing on [1, 2],
5
hence f (x) ≥ f (1) = 0.

2 2
a5b + b5a ≤ 2.
Solution. Assume that a ≥ b. For a = 2 and b = 0, the inequality is obvious. Otherwise,

using the substitution a = 1 + x and b = 1 − x, 0 ≤ x < 1, we can write the desired
inequality as
2 2
e5(1−x) ln(1+x) + e5(1+x) ln(1−x) ≤ 2.
According to Lemma below, it suffices to show that f (x) ≤ 2, where
L
f (x) = e5(u−v) + e−5(u+v) ,
.M
7 3 31 5 5 17 4 9 6
u= x+ x + x , v = x2 + x + x .
D
3 30 2 12 20
If f 0 (x) ≤ 0, then f (x) is decreasing, hence f (x) ≤ f (0) = 2. Since
A
PI
f 0 (x) = 5(u0 − v 0 )e5(u−v) − 5(u0 + v 0 )e−5(u+v) ,
31 4 17 3 27 5
M
u0 = 1 + 7x 2 + x , v 0 = 5x + x + x ,
6 3 10
LY
the inequality f 0 (x) ≤ 0 becomes
e−10u (u0 + v 0 ) ≥ u0 − v 0
O
For the nontrivial case u0 − v 0 > 0, we rewrite this inequality as g(x) ≥ 0, where
.M
g(x) = −10u + ln(u0 + v 0 ) − ln(u0 − v 0 ).

W
If g 0 (x) ≥ 0, then g(x) is increasing, hence g(x) ≥ f (0) = 2. We have

W
u00 + v 00 u00 − v 00
g 0 (x) = −10u0 + − 0 ,
u0 + v 0 u − v0
W
where
62 3 27 4
u00 = 14x + x , v 00 = 5 + 17x 2 + x .
3 2
Thus, the inequality g 0 (x) ≥ 0 is equivalent to
u0 v 00 − v 0 u00 ≥ 5u0 (u0 + v 0 )(u0 − v 0 ),
a1 t + a2 t 2 + a3 t 3 + a4 t 4 + a5 t 5 + a6 t 6 + a7 t 7 ≥ 0,
where t = x 2 , 0 ≤ t < 1, and
a1 = 2, a2 = 321.5, a3 ≈ 152.1, a4 ≈ −498.2,

a5 ≈ −168.5, a6 ≈ 356.0, a7 ≈ 188.3.

300t 2 + 150t 3 − 500t 4 − 200t 5 + 250t 6 ≥ 0.
Since the last inequality is equivalent to the obvious inequality
50t 2 (1 − t)(6 + 9t − t 2 − 5t 3 ) ≥ 0,
the proof is completed. The equality holds for a = b = 1.

Lemma. If −1 < t < 1, then
L
5 2 7 3 17 4 31 5 9 6
.M
(1 − t)2 ln(1 + t) ≤ t − t + t − t + t − t .
2 3 12 30 20
Proof. We shall show that
D
A
1 1 1 1

(1 − t)2 ln(1 + t) ≤ (1 − t)2 t − t 2 + t 3 − t 4 + t 5
2 3 4 5 PI
5 2 7 3 17 4 31 5 9 6
≤t− t + t − t + t − t .
M
2 3 12 30 20
The left inequality is equivalent to f (t) ≥ 0, where
LY
1 2 1 3 1 4 1 5
f (t) = t − t + t − t + t − ln(1 + t).
O
2 3 4 5
.M
Since
t5
f 0 (t) = ,
1+ t
W
f (t) is decreasing on (−1, 0] and increasing on [0, 1); therefore, f (t) ≥ f (0) = 0.
The right inequality is equivalent to t 6 (t − 1) ≤ 0, which is clearly true.
W
W

p p
a2 b
+ b2 a
≤ 2.
Solution. Assume that a ≥ b. For a = 2 and b = 0, the inequality is obvious. Otherwise,

using the substitution a = 1 + x and b = 1 − x, 0 ≤ x < 1, we can write the desired
inequality as f (x) ≤ 2, where
p p
f (x) = e21−x ln(1+x)
+ e2 1+x ln(1−x)
.
There are two cases to consider.

13
Case 1: ≤ x < 1. If f 0 (x) ≤ 0, then f (x) is decreasing, and hence
20
q q
7 33 5/4 2
13 33 7 5 1
5
5
f (x) ≤ f = + < + < 2.
20 20 20 3 4
Since
p
2 1 − x ln(1 + x) 2p1−x ln(1+x)

0
f (x) = − p e
1+ x 1− x
p
2 1 + x ln(1 − x) 2p1+x ln(1−x)

L
− − p e
1− x 1+ x
.M
p
2 1 − x ln(1 + x) 2p1−x ln(1+x)

< − p e ,
1+ x 1− x
D
A
it is enough to show that g(x) ≤ 0,where
g(x) =
2(1 − x)
1+ x
− ln(1 + x).
PI
M
Clearly, g is decreasing, and hence
LY
13 14 33

g(x) ≤ g = − ln < 0.
20 33 20
O
13
.M
Case 2: 0 ≤ x ≤ . According to Lemma below, it suffices to show that f (x) ≤ 2,

20
where
11 3 1 4
2
+ 12 x −2 x 2
+ 11 3 1 4
12 x + 2 x )
f (x) = e2x−2x + e−(2x+2x .
W
If f 0 (x) ≤ 0, then f (x) is decreasing, hence f (x) ≤ f (0) = 2. Since

W
11 2 2 11 3 1 4
f 0 (x) = (2 − 4x + x − 2x 3 )e2x−2x + 12 x − 2 x
W
4
11 2 2 11 3 1 4
−(2 + 4x + x + 2x 3 )e−(2x+2x + 12 x + 2 x ) ,
4
0
the inequality f (x) ≤ 0 is equivalent to
11
x3 8 − 16x + 11x 2 − 8x 3
e−4x− 6 ≥ .
8 + 16x + 11x 2 + 8x 3
For the non-trivial case 8 − 16x + 11x 2 − 8x 3 > 0, rewrite this inequality as g(x) ≥ 0,
where
11 3
g(x) = −4x − x − ln(8 − 16x + 11x 2 − 8x 3 ) + ln(8 + 16x + 11x 2 + 8x 3 ).
6
If g 0 (x) ≥ 0, then g(x) is increasing, hence g(x) ≥ g(0) = 0. From
11 2 (16 + 24x 2 ) − 22x (16 + 24x 2 ) + 22x

g 0 (x) = −4 − x + + ,
2 8 + 11x 2 − (16x + 8x 3 ) 8 + 11x 2 + (16x + 8x 3 )
it follows that g 0 (x) ≥ 0 is equivalent to
1
(16 + 24x 2 )(8 + 11x 2 ) − 22x(16x + 8x 3 ) ≥ (8 + 11x 2 )[(8 + 11x 2 )2 − (16x + 8x 3 )2 ].
4
Since
L
(8 + 11x 2 )2 − (16x + 8x 3 )2 ≤ (8 + 11x 2 )2 − 256x 2 − 256x 4 ≤ 16(4 − 5x 2 ),
.M
D
(4 + 6x 2 )(8 + 11x 2 ) − 11x(8x + 4x 3 ) ≥ (8 + 11x 2 )(4 − 5x 2 ).
A
This reduces to 77x 4 ≥ 0. The proof is completed. The equality holds for a = b = 1.
Lemma. If −1 < t ≤
13
, then
PI
M
20
p 11 3 1 4
LY
1 − t ln(1 + t) ≤ t − t 2 + t − t .
24 4
O
Proof. We consider two cases.

13
.M
Case 1: 0 ≤ t ≤ . We can prove the desired inequality by multiplying the following

20
inequalities
p 1 1 1 3
W
1 − t ≤ 1 − t − t2 − t ,
2 8 16
W
1 1 1 1
ln(1 + t) ≤ t − t 2 + t 3 − t 4 + t 5 ,
2 3 4 5
W
1 1 1 3 1 1 1 1 11 3 1 4

1 − t − t2 − t t − t2 + t3 − t4 + t5 ≤ t − t2 + t − t .
2 8 16 2 3 4 5 24 4
The first inequality is equivalent to f (t) ≥ 0, where
1 1 2 1 3 1

f (t) = ln 1 − t − t − t − ln(1 − t).
2 8 16 2
Since
1 8 + 4t + 3t 2 5t 3
f 0 (t) = − = ≥ 0,
2(1 − t) 16 − 8t − 2t 2 − t 3 2(1 − t)(16 − 8t − 2t 2 − t 3 )
f (t) is increasing, and hence f (t) ≥ f (0) = 0.

The second inequality is equivalent to f (t) ≥ 0, where
1 2 1 3 1 4 1 5
f (t) = t − t + t − t + t − ln(1 + t).
2 3 4 5
Since
1 t5
f 0 (t) = 1 − t + t 2 − t 3 + t 4 − = ≥ 0,
1+ t 1+ t
f (t) is increasing, and hence f (t) ≥ f (0) = 0.
The third inequality is equivalent to
L
t 4 (160 − 302t + 86t 2 + 9t 3 + 12t 4 ) ≥ 0.
.M
This is true since
D
160 − 302t + 86t 2 + 9t 3 + 12t 4 ≥ 2(80 − 151t + 43t 2 ) > 0.
A
Case 2: −1 < t ≤ 0. Write the desired inequality as
PI
M
p 11 3 1 4
− 1 − t ln(1 + t) ≥ −t + t 2 − t + t .
LY
24 4
This is true if
O
p 1 1
1 − t ≥ 1 − t − t 2,
2 8
.M
1 1
− ln(1 + t) ≥ −t + t 2 − t 3 + t 4 ,
3 4
W
1 1 1 1 11 3 1 4

1 − t − t 2 −t + t 2 − t 3 + t 4 ≥ −t + t 2 − t + t .
2 8 3 4 24 4
W
The first inequality is equivalent to f (t) ≥ 0, where

W
1 1 1 2

f (t) = ln(1 − t) − ln 1 − t − t .
2 2 8
Since
−1 2(2 + t) −3t 2
f 0 (t) = + = ≤ 0,
2(1 − t) 8 − 4t − t 2 2(1 − t)(8 − 4t − t 2 )
f (t) is decreasing, and hence f (t) ≥ f (0) = 0.
The second inequality is equivalent to f (t) ≥ 0, where
1 2 1 3 1 4
f (t) = t − t + t − t − ln(1 + t).
2 3 4
Since
1 −t 4
f 0 (t) = 1 − t + t 2 − t 3 − = ≤ 0,
1+ t 1+ t
f (t) is decreasing, and hence f (t) ≥ f (0) = 0.
The third inequality reduces to the obvious inequality
t 4 (10 − 8t − 3t 2 ) ≥ 0.
L
.M
a b(1 − a b)2 a b(1 − a b)2
≤ a b+1 + b a+1 − 2 ≤ .
D
2 3
A
PI
Solution. Assume that a ≥ b, which yields 1 ≤ a ≤ 2 and 0 ≤ b ≤ 1.
M
(a) To prove the left inequality we apply Lemma 1 below. For x = a and k = b, we
have
LY
b(1 + b) b(1 + b)(1 − b)

a b+1 ≥ 1 + (1 + b)(a − 1) + (a − 1)2 − (a − 1)3 ,
O
2 6
.M
b(1 + b) b(1 + b)
a b+1 ≥ a − b + a b + (a − 1)2 − (a − 1)4 . (*)
2 6
Also, for x = b and k = a − 1, we have
W
a(a − 1) a(a − 1)(2 − a)

b a ≥ 1 + a(b − 1) + (b − 1)2 − (b − 1)3 ,
W
2 6
W
a ab
b a ≥ 1 − a + a b + (a − 1)3 + (a − 1)4 ,
2 6
ab a b2
b a+1 ≥ b − a b + a b2 + (a − 1)3 + (a − 1)4 . (**)
2 6
Summing up (*) and (**) gives
b(3 − a b) b(1 + b − a b)
a b+1 + b a+1 − 2 ≥ −b(a − 1)2 + (a − 1)2 − (a − 1)4 .
2 6
Since
b(3 − a b) b
−b(a − 1)2 + (a − 1)2 = (a − 1)4 ,
2 2
we have
b b(1 + b − a b)
a b+1 + b a+1 − 2 ≥ (a − 1)4 − (a − 1)4
2 6
a b(1 + b) ab a b(1 − a b)2
= (a − 1)4 ≥ (a − 1)4 = .
6 6 6
(b) To prove the right inequality we apply Lemma 2 below. For x = a and k = b,
we have
(b + 1)b (b + 1)b(b − 1)
a b+1 ≤ 1 + (b + 1)(a − 1) + (a − 1)2 + (a − 1)3
L
2 6
.M
(b + 1)b(b − 1)(b − 2)
+ (a − 1)4 ,
24
D
b(b + 1) b(b + 1) a b(b + 1)
a b+1 ≤ 1 + (b + 1)(a − 1) + (a − 1)2 − (a − 1)4 + (a − 1)5 .
A
2 6 24
Also, for x = b and k = a, we have PI
M
a(a + 1) a(a + 1) a b(a + 1)
b a+1 ≤ 1 + (a + 1)(b − 1) + (b − 1)2 − (b − 1)4 + (b − 1)5 .
2 6 24
LY
Summing these inequalities gives

O
a2 + b2 + 2 a2 + b2 + 2 ab
a b+1 + b a+1 − 2 ≤ −2(a − 1)2 + (a − 1)2 − (a − 1)4 − (a − 1)6
.M
2 6 12
a2 + b2 − 2 a2 + b2 + 2 a2 + b2 + 2
≤ (a − 1)2 − (a − 1)4 = (a − 1)4 − (a − 1)4
W
2 6 6
ab a b(1 − a b)2
W
= (a − 1)4 = .
3 3
W

Lemma 1. If x ≥ 0 and 0 ≤ k ≤ 1, then
k(1 + k) k(1 + k)(1 − k)

x k+1 ≥ 1 + (1 + k)(x − 1) + (x − 1)2 − (x − 1)3 ,
2 6
with equality for x = 1, for k = 0, for k = 1.
Proof. For k = 0 and k = 1, the inequality is an identity. For fixed k, 0 < k < 1, let us
define
k(1 + k) k(1 + k)(1 − k)
f (x) = x k+1 − 1 − (1 + k)(x − 1) − (x − 1)2 + (x − 1)3 .
2 6
We need to show that f (x) ≥ 0. We have
1 k(1 − k)
f 0 (x) = x k − 1 − k(x − 1) + (x − 1)2 ,
1+k 2
1
f 00 (x) = x k−1 − 1 + (1 − k)(x − 1),
k(1 + k)
1
f 000 (x) = −x k−2 + 1.
k(1 + k)(1 − k)
Case 1: 0 ≤ x ≤ 1. Since f 000 ≤ 0, f 00 is decreasing, f 00 (x) ≥ f 00 (1) = 0, f 0 is increasing,
L
f 0 (x) ≤ f 0 (1) = 0, f is decreasing, f (x) ≥ f (1) = 0.
.M
Case 2: x ≥ 1. Since f 000 ≥ 0, f 00 is increasing, f 00 (x) ≥ f 00 (1) = 0, f 0 is increasing,
f 0 (x) ≥ f 0 (1) = 0, f is increasing, f (x) ≥ f (1) = 0.
D
Lemma 2. If either x ≥ 1 and 0 ≤ k ≤ 1, or 0 ≤ x ≤ 1 and 1 ≤ k ≤ 2, then
A
(k + 1)k (k + 1)k(k − 1)
x k+1 ≤ 1 + (k + 1)(x − 1) +
2
PI
(x − 1)2 +
6
(x − 1)3
M
(k + 1)k(k − 1)(k − 2)
+ (x − 1)4 ,
24
LY
with equality for x = 1, for k = 0, for k = 1, for k = 2.

O
Proof. For k = 0, k = 1 and k = 2, the inequality is an identity. For fixed k, k ∈

(0, 1) ∪ (1, 2), let us define
.M
(k + 1)k (k + 1)k(k − 1)
f (x) = x k+1 − 1 − (k + 1)(x − 1) − (x − 1)2 − (x − 1)3
2 6
W
(k + 1)k(k − 1)(k − 2)
− (x − 1)4 .
W
24
We need to show that f (x) ≤ 0. We have
W
1 k(k − 1) k(k − 1)(k − 2)

f 0 (x) = x k − 1 − k(x − 1) − (x − 1)2 − (x − 1)3 ,
k+1 2 6
1 (k − 1)(k − 2)
f 00 (x) = x k−1 − 1 − (k − 1)(x − 1) − (x − 1)2 ,
k(k + 1) 2
1
f 000 (x) = x k−2 − 1 − (k − 2)(x − 1),
k(k + 1)(k − 1)
1
f (4) (x) = x k−3 − 1.
k(k + 1)(k − 1)(k − 2)
Case 1: x ≥ 1, 0 < k < 1. Since f (4) (x) ≤ 0, f 000 (x) is decreasing, f 000 (x) ≤ f 000 (1) =
0, f 00 is decreasing, f 00 (x) ≤ f 00 (1) = 0, f 0 is decreasing, f 0 (x) ≤ f 0 (1) = 0, f is
decreasing, f (x) ≤ f (1) = 0.
Case 2: 0 ≤ x ≤ 1, 1 < k < 2. Since f (4) ≤ 0, f 000 is decreasing, f 000 (x) ≥ f 000 (1) = 0, f 00
is increasing, f 00 (x) ≤ f 00 (1) = 0, f 0 is decreasing, f 0 (x) ≥ f 0 (1) = 0, f is increasing,
f (x) ≤ f (1) = 0.
L
.M
a2b + b2a ≤ 1.
D
Solution. Without loss of generality, assume that
A
0≤ b≤
1
2
≤ a ≤ 1.
PI
M
Applying Lemma 1 below for c = 2b, 0 ≤ c ≤ 1, we get
LY
a2b ≤ (1 − 2b)2 + 4a b(1 − b) − 2a b(1 − 2b) ln a,

O
a2b ≤ 1 − 4a b2 − 2a b(a − b) ln a.
.M
Similarly, applying Lemma 2 below for d = 2a − 1, d ≥ 0, we get

W
b2a−1 ≤ 4a(1 − a) + 2a(2a − 1) ln(2a + b − 1),

W
b2a ≤ 4a b2 + 2a b(a − b) ln a.
W
Adding up these inequalities, the desired inequality follows. The equality holds for
a = b = 1/2, for a = 0 and b = 1, and for a = 1 and b = 0.
Lemma 1. If 0 < a ≤ 1 and c ≥ 0, then
a c ≤ (1 − c)2 + ac(2 − c) − ac(1 − c) ln a,
with equality for a = 1, for c = 0, and for c = 1.

Proof. Making the substitution a = e−x , x ≥ 0, we need to prove that f (x) ≥ 0, where
f (x) = (1 − c)2 e x + c(2 − c) + c(1 − c)x − e(1−c)x ,

f 0 (x) = (1 − c)[(1 − c)e x + c − e(1−c)x ].

If f 0 ≥ 0 on [0, ∞), then f is increasing, and hence f (x) ≥ f (0) = 0. In order to prove
that f 0 ≥ 0, we consider two cases.
Case 1: 0 ≤ c ≤ 1. By the weighted AM-GM inequality, we have
(1 − c)e x + c ≥ e(1−c)x ,
and hence f 0 (x) ≥ 0.

Case 2: c ≥ 1. By the weighted AM-GM inequality, we have
(c − 1)e x + e(1−c)x ≥ c,
L
.M
which yields
f 0 (x) = (c − 1)[(c − 1)e x + e(1−c)x − c] ≥ 0.
D
A
Lemma 2. If 0 ≤ b ≤ 1 and d ≥ 0, then
PI
b d ≤ 1 − d 2 + d(1 + d) ln(b + d),
M
with equality for d = 0, and for b = 0 and d = 1.
LY
Proof. Write the inequality as

O
(1 + d)[1 − d + d ln(b + d)] ≥ b d .

.M
Excepting the equality cases, since
1 − d + d ln(b + d) ≥ 1 − d + d ln d > 0,
W
we can rewrite the inequality in the form

W
ln(1 + d) + ln[1 − d + d ln(b + d)] ≥ d ln b.

W
Using the substitution b = e−x − d, where − ln(1 + d) ≤ x ≤ − ln d, we need to prove

that f (x) ≥ 0, where
f (x) = ln(1 + d) + ln(1 − d − d x) + d x − d ln(1 − d e x ).
Since
d 2 (e x − 1 − x)
f 0 (x) = ≥ 0,
(1 − d − d x)(1 − d e x )
f is increasing, and hence
f (x) ≥ f (− ln(1 + d)) = ln[1 − d 2 + d(1 + d) ln(1 + d)].

To complete the proof, we only need to show that −d 2 + d(1 + d) ln(1 + d) ≥ 0; that is,
(1 + d) ln(1 + d) ≥ d.
−d
This inequality follows from e x ≥ 1 + x, where x = .
1+d
Conjecture. If a, b are nonnegative real numbers such that 1 ≤ a + b ≤ 15, then
a2b + b2a ≤ a a+b + b a+b .
L
.M
D
A
2a a b b ≥ a2b + b2a .
PI
M
Solution. Taking into account the inequality a2b + b2a ≤ 1 in the preceding P 3.32, it
LY

2a a b b ≥ 1.
O

.M
2a a b b ≥ a a+b + b a+b ,
a b b a
2≥ + .
W
b a
Since a < 1 and b < 1, we apply Bernoulli’s inequality as follows
W
a b a
b a b
W
+ ≤1+ b −1 +1+a − 1 = 2.
b a b a
Thus, the proof is completed. The equality holds for a = b = 1/2.
a−2a + b−2b ≤ 4.
Solution. Applying Lemma below, we have

a−2a ≤ 4 − 2 ln 2 − 4(1 − ln 2)a,
b−2b ≤ 4 − 2 ln 2 − 4(1 − ln 2)b.
Adding these inequalities, the desired inequality follows. The equality holds for a = b =
1/2.
Lemma. If x ∈ (0, 1], then
x −2x ≤ 4 − 2 ln 2 − 4(1 − ln 2)x,
with equality for x = 1/2.
L
Proof. Write the inequality as
.M
1 −2x 1
x ≤ 1 − c − (1 − 2c)x, c= ln 2 ≈ 0.346.
4 2
D
This is true if f (x) ≤ 0, where
A
f (x) = −2 ln 2 − 2x ln x − ln[1 − c − (1 − 2c)x].
We have
PI
1 − 2c
M
f 0 (x) = −2 − 2 ln x + ,
1 − c − (1 − 2c)x
LY
2 (1 − 2c)2 g(x)
f 00 (x) = − + = ,
x [1 − c − (1 − 2c)x]2 x[1 − c − (1 − 2c)x]2
O
where
g(x) = 2(1 − 2c)2 x 2 − (1 − 2c)(5 − 6c)x + 2(1 − c)2 .
.M
Since
g 0 (x) = (1 − 2c)[4(1 − 2c)x − 5 + 6c] ≤ (1 − 2c)[4(1 − 2c) − 5 + 6c]
W
= (1 − 2c)(−1 − 2c) < 0,

W
g is decreasing on (0, 1], hence g(x) ≥ g(1) = −2c 2 + 4c − 1 > 0, f 00 (x) > 0 for
x ∈ (0, 1]. Since f 0 is increasing and f 0 (1/2) = 0, we have f 0 (x) ≤ 0 for x ∈ (0, 1/2],
W
and f 0 (x) ≥ 0 for x ∈ [1/2, 1]. Therefore, f is decreasing on (0, 1/2] and increasing on
[1/2, 1], hence f (x) ≥ f (1/2) = 0.
Remark. According to the inequalities in P 3.32 and P 3.34, the following inequality
holds for all positive numbers a, b such that a + b = 1:
1 1

a2b + b2a + ≤ 4.
a2a b2b
Actually, this inequality holds for all a, b ∈ (0, 1]. In this case, it is sharper than the
inequality in P 3.15.
L
D
.M
A
PI
M
LY
O
.M
W
W
W
Chapter 4
Bibliography
L
D
.M
A
PI
M
LY
O
.M
W
W
W
473

(Vasile Cîrtoaje) Mathematical Inequalities Vol 2

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

(Vasile Cîrtoaje) Mathematical Inequalities Vol 2

Transféré par

Droits d'auteur :

Formats disponibles

Vasile Cîrtoaje

RATIONAL AND NONRATIONAL

UNIVERSITY OF PLLOIESTI, ROMANIA

3.2 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425

Symmetric Rational Inequalities

1.2. If a, b, c are nonnegative real numbers, then

1.3. If a, b, c are positive real numbers, then

a(b + c) b(c + a) c(a + b)

a(b + c) b(c + a) c(a + b)

1.9. Let a, b, c be positive real numbers. Prove that

1.10. Let a, b, c be positive real numbers. Prove that

1.13. Let a, b, c be positive real numbers. Prove that

1.16. If a, b, c are positive real numbers, then

(a2 + b2 )(a2 + c 2 ) (b2 + c 2 )(b2 + a2 ) (c 2 + a2 )(c 2 + b2 )

1.19. Let a, b, c be positive real numbers such that a + b + c = 3. Prove that

1.22. Let a, b, c be positive real numbers such that a + b + c = 3. Prove that

1.23. Let a, b, c be positive real numbers such that a + b + c = 3. Prove that

1.24. Let a, b, c be nonnegative real numbers such that a + b + c = 2. Prove that

1.25. Let a, b, c be nonnegative real numbers such that a + b + c = 1. Prove that

1.26. Let a, b, c be positive real numbers such that a + b + c = 1. Prove that

1.27. Let a, b, c be positive real numbers such that a + b + c = 3. Prove that

1.30. Let a, b, c be nonnegative real numbers such that a + b + c = 3. Prove that

1.31. Let a, b, c be nonnegative real numbers such that a ≥ b ≥ 1 ≥ c and a + b + c = 3.

1.32. Let a, b, c be nonnegative real numbers such that a + b + c = 3. Prove that

1.33. Let a, b, c be nonnegative real numbers such that a ≥ 1 ≥ b ≥ c and a + b + c = 3.

1.34. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. Prove that

1.35. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. Prove that

1.38. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. If k ≥ 2 + 3,

a+k b+k c+k 1+k

1.39. Let a, b, c be nonnegative real numbers such that a2 + b2 + c 2 = 3. Prove that

a(b + c) b(c + a) c(a + b)

1.40. Let a, b, c be positive real numbers such that a2 + b2 + c 2 = 3. Prove that

1.41. Let a, b, c be positive real numbers such that a2 + b2 + c 2 = 3. Prove that

1.42. Let a, b, c be positive real numbers such that a2 + b2 + c 2 = 3. Prove that

1.45. Let a, b, c be positive real numbers such that a bc = 1. Prove that

1.46. Let a, b, c be positive real numbers such that a bc = 1. Prove that

1.47. Let a, b, c be positive real numbers such that a bc = 1. Prove that

1.48. Let a, b, c be positive real numbers such that a bc = 1. Prove that

1.49. Let a, b, c be positive real numbers such that a bc = 1. Prove that

1.52. Let a, b, c be nonnegative real numbers such that

7(a2 + b2 + c 2 ) = 11(a b + bc + ca).

a(2a + b + c) b(2b + c + a) c(2c + a + b)

a2 (b + c)2 b2 (c + a)2 c 2 (a + b)2

a+b b+c c+a 2(a + b + c)

3a3 + 13a bc 3b3 + 13a bc 3c 3 + 13a bc

2a2 + bc 2b2 + ca 2c 2 + a b 9(a2 + b2 + c 2 )

a(b + c) b(c + a) c(a + b)

1.73. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. Prove that

1.74. Let a, b, c be nonnegative real numbers such that a b + bc + ca = 3. Prove that

2a2 + 3bc 2b2 + 3ca 2c 2 + 3a b

2a2 + 5bc 2b2 + 5ca 2c 2 + 5a b 21

a2 + 128bc b2 + 128ca c 2 + 128a b

a2 + 64bc b2 + 64ca c 2 + 64a b

1.92. If a, b, c are nonnegative real numbers, then

1.93. Let a, b, c be positive real numbers. If k > 0, then

3a2 − 2bc 3b2 − 2ca 3c 2 − 2a b 3

a(b + c) b(c + a) c(a + b) a b + bc + ca

a(b + c) b(c + a) c(a + b) a2 + b2 + c 2

a(b + c) b(c + a) c(a + b) (a + b + c)2

1.105. If a, b, c are the lengths of the sides of a triangle, then

a(b + c) b(c + a) c(a + b)

1.106. If a, b, c are real numbers, then

1.107. If a, b, c are nonnegative real numbers, then