Vous êtes sur la page 1sur 3

Useful Inequalities

{x2 > 0}


n k
k

binomial

v0.26a April 26, 2014

Cauchy-Schwarz

Minkowski

H
older




xi y i

n
P

|xi + yi |p

i=1

i=1

n
P

i=1

Bernoulli

2

n
P

|xi yi |

(1 +

x)r

n
P

i=1

1

n
P

i=1

x2i



n
P

|xi |p

1 + rx

i=1

i=1

(1 +

1/p 

1

n
P

i=1

|yi |q

n
P

i=1

|yi |p

1/q

1

1
p

1
q

for x [1,

1
),
r1

r > 1.

e x xe


x n
n

1
2x

1
<

1+
xn
n!

for x R, and

ex 1 + x +
ex

1 + x,

xx

x
2

<

x2
2


x n
n

+ 1 ex 1 +

for x 0, reverse for x 0.

for x [0, 1.59] and

x2

x2
n

ex 1

x+1

2x

2 x+12 x <

Stirling

means

min{xi }

power means

for n > 1, |x| n.


x n+x/2
n

x
2

for x (0, 1).

x1/r (x 1) rx(x1/r 1)

for x, r 1.
xy
y
y
x
x
> e x+y
x + y > 1 and e > 1 + x
y
x2 1
x2 +1

x1
x

2x
2+x

ln(1 + x)

ln(1 + x) x

x2
2

ln(1 x) x
x3
2

1
n

hyperbolic

x cos x
2
x

x3
sinh2 x

x
2

x cos x
1x2 /3

Jensen

x cos2 (x/2) sin x (x cos x + 2x)/3

sin x x cos(x/2) x x +

cosh(x) + sinh(x)

Chebyshev

ex(+x/2)

x3
3

tan x

sin x,

x2
,
sinh x

for x R, [1, 1].

2n

and

1
x

i=0

<

1.

xi

n
i

2n

n
2

for

for n d 0.

d 0.

for (0, 21 ).

x+1


n n 1/(12n+1)
e
e

n
1
i xi

wi |xi |p

wi |xi |p1

xy
xy

x1 < 2 x2 x1

n!

1/n

x+
2

P
wi |xi |q
P i
q1
i wi |xi |

(xy) 4

x1 y +x y 1
2

Sk 2 Sk1 Sk+1
P
1
Sk =  n

1
n

2n


n n 1/12n
e
e

xi

1
n

for x 1.

en


n n
e

xi 2 max{xi }

for p q, wi 0.

xy
ln(x)ln(y)
x+y
2

p
k

and

1i1 <<ik n

x+
2

2

x+y
2

for x, y > 0.

for x, y > 0, [0, 1].

Sk

(k+1)

ai 1 ai 2 aik ,

 P
pi xi i pi (xi )

n
P

f (ai )g(bi )pi

n
P

ai b i

i=1


all for x 0,

for n > k > 0.

2n

Sk+1

and

where pi 0,

for 1 k < n,

ai 0.
P

pi = 1, and convex.

n
P

f (ai )pi

i=1

n
 P

i=1

n

P
g(bi )pi
f (ai )g(bni+1 )pi
i=1

for a1 an , b1 bn and f, g nondecreasing, pi 0,





 

Alternatively: E f (X)g(X) E f (X) E g(X) .

for x [0, 0.43], reverse elsewhere.


x

Alternatively: (E [X]) E [(X)]. For concave the reverse holds.

for x [0, 0.45], reverse elsewhere.

x 3 cos x x x3 /6 x cos

x 1, ln(x) n(x n 1) for x, n > 0.

for x 0, reverse for x (1, 0].


Pn 1
i=1 i ln(n) + 1

x3
4

x
2

x cos x

for x, y R.

Heinz

for x, y > 0.
2

2n


P
P
p 1/p , w 0,
Mp Mq for p q, where Mp =
i
i wi |xi |
i wi = 1.
Q
In the limit M0 = i |xi |wi , M = mini {xi }, M = maxi {xi }.

x
x+1

ln(n + 1) < ln(n) +

trigonometric

x2 1
2x

ln(x)


n n
e

log mean

Maclaurin

2 y exy 1 + x y + exy , and ex x + ex

logarithm

Lehmer

for x, n > 0.

for x [0, 1].

+1

en d

square root

(iv) q < 0 < p , q > x > 0, (v) q < 0 < p , p < x < 0.
ex 1 +

nn
kk (nk)nk

Pd

n
for n d
i=0 i
d

Pd
n
n
d
i=0 i d 1 + n2d+1
P
n
n 
n 
1
n
i=0 i 12 n
n

for x R, n N.

(ii) p < q < x < 0, (iii) q > p > x > 0. Reverse for:

Pd

= 1.

(a + b)n an + nb(a + b)n1 for a, b 0, n N.


p
q
1+ x
1 + xq
for (i) x > 0, p > q > 0,
p

exponential

n
k

2n(1)

for x [1, 0], n N.

(1 + nx)n+1 (1 + (n + 1)x)n

nd

Pd

for x [0, 1], r R \ (0, 1).

rx
1(r1)x

n
i=0 i


en k
,
k

1
1
n k 0 and 4n (1 8n
) n 4n (1 9n
).

+n2
for n1 k1 0, n2 k2 0.
nk1 +k
k1 k2
1
2

nH()

n 
G n G for G = 2
, H(x) = log2 (xx (1x)1x ).
2

for p 1.

for p, q > 1,

for

k

n1 n2 

nk
k!

n

n
4k!

for x 1, r R \ (0, 1). Reverse for r [0, 1].

1
1nx

(1 + x)r 1 +

yi2

|xi |p

(1 + x)r 1 + (2r 1)x


x)n

n
P

n
k

rearrangement

i=1

n
P

i=1

ai b(i)

n
P

ai bni+1

i=1

for a1 an ,

b1 bn and a permutation of [n]. More generally:


n
n
n
P
P
P
fi (bi )
fi (b(i) )
fi (bni+1 )
i=1

i=1

i=1


with fi+1 (x) fi (x) nondecreasing for all 1 i < n.

pi = 1.

Weierstrass

1 xi

 wi

w i xi

where xi 1, and

either wi 1 (for all i) or wi 0 (for all i).


P
If wi [0, 1],
wi 1 and xi 1, the reverse holds.
Young

Kantorovich

( px1p
P

1 1
)
qxq

xi

xi
yi

0<m
sum-integral

RU

Cauchy

(a)

L1

Hermite

Chong

n
P

P
Shapiro

i
i

i=L

n
ai
bi

ai

Rb
a

2 P

n
Q

i=1

xi yi

R U +1
L

1
q

product diff.

Q n

ai

= 1.
sum-product

a
b

n
2

f (x) dx

(a)+(b)
2

Karamata

n
Q

ai a i

Muirhead
ai a(i)

for ai > 0.

Hadamard
Schur

i=1 j=1

Pn

2
i=1 i

2
i,j=1 Aij

Pk

and

i=1 di

Pk

i=1 i

Hilbert

Hardy

Pn
a
P
xi i
i=1 ai xi
Qn
P

for xi [0, 12 ], ai [0, 1],


ai = 1.
n
ai
i=1 (1 xi )
i=1 ai (1 xi )
Qn

i=1

a1 b 1

given that
Mahler

n
Q

Pn
a21

xi + yi

i=1

Abel
Milne

b1 min
k

Pn

ai b i
Pn

i=2

k
P

i=1

i=1 (ai

>

+ bi )

a21

2
i=2 ai

1/n

ai

2

n
P

i=1

n
Q

i=1

or

1/n
xi

b21

Pn
>

2 b2
1
i=2 ai
Pn
2
i=2 bi .

n
Q

i=1

ai bi b1 max

 P

Pn

i=1

m Q
n
P

|ai bi |

ai(j)

j=1 i=1

a1+x
b1x
i
i

Pn

i=1

(ai )

ai bi
n
i=1 ai +bi

1/n
yi

Pn

2
i=2 bi

1
n!

 P

for |ai |, |bi | 1.

m P
n
Q

and

j=1 i=1

aij

m P
n
Q

j=1 i=1

ai(j)

 P
 P

1+y 1y
1y 1+y
a1x
b1+x

bi
bi
i ai
i ai
i
i

Pn

(bi )

i=1

n
1

xa
xa
(n)
(1)

m=1

for a1 a2 an and b1 bn ,
Pt
Pt
i=1 ai
i=1 bi for all 1 t n,

an

Mathieu

1
c2 +1/2

<

Copson

 P
P

n=1

n=1

Kraft
LYM

kn

XA

Sauer-Shelah

Pn

i=1

for b1 bn 0.
ai

 Pn

i=1 bi

n
1
xb(1)
xb(n)

m=1

a2m

 1 P
2

n=1

a2n

2n
n=1 (n2 +c2 )2

ak
k

p

n 1
|X|

pp

n=1

1
c2

<

an p

n=1

n2 a2n

1

2 2
n=1 bn

for am , bn R.

for an 0, p > 1, reverse if p (0, 1).

for c(i) depth of leaf i of binary tree, sum over all leaves.
1, A 2[n] , no set in A is subset of another set in A.

|A| |str(A)|

Pr
Pr

vc(A)
P
i=0

n
i

for A 2[n] , and


vc(A) = max{|X| : X str(A)}.

n
 W

k

P
Ai
(1)j1 Sj

for 1 k n, k odd,

n
 W

k

P
Ai
(1)j1 Sj

for 2 k n, k even.

i=1

Sk =

for an R.

for c 6= 0.

str(A) = {X [n] : X shattered by A},


Bonferroni

ai

4

2c(i) 1

where xi , yi > 0.

i=1

1
n!

With max{m, n} instead of m + n, we have 4 instead of .


P  a1 +a2 ++an p  p p P
p
p1
for an 0, p > 1.
n=1
n=1 an
n

i=1

k
P

am bn
n=1 m+n

1 n the eigenvalues, d1 dn the diagonal elements.

Acz
el

P

Carlson

for 1 k n.

A is an n n matrix. For the second inequality A is symmetric.

Ky Fan

aij

|ak |

k=1

xi 0 and the sums extend over all permutations of [n].

where A is an n n matrix.

Pn

i=1 bi

Pn

where a1 a2 an and b1 b2 bn and {ak }  {bk },

i=1

where xi > 0, (xn+1 , xn+2 ) := (x1 , x2 ),

A2ij

Qn

with equality for t = n and is convex (for concave the reverse holds).

for convex.

where x, y, z 0, t > 0
(det A)2

1/k

and {ai }  {bi } (majorization), i.e.

xt (x y)(x z) + y t (y z)(y x) + z t (z x)(z y) 0


n P
n
Q

m Q
n
P

|ai |

for 1 x y 0, and i = 1, . . . , n.

and n 12 if even, n 23 if odd.


Schur

i=1

k
i=1

where 0 ai1 aim for i = 1, . . . , n and is a permutation of [n].

for f nondecreasing.

for ai , bi 0, or more generally:


P
P
for concave, and a :=
ai , b :=
bi .

k=1

j=1 i=1

2

where a < b, and convex.

(x) dx

b
a

for x, y 0, p, q > 0,

1
p

Q

for xi , yi > 0,

A = (m + M )/2, G = mM .

f (i)

and

(b)

a log

bi 

A
G

yq
q

Pn

Callebaut

PU

1
ba

xi
xi+1 +xi+2

i=1

M < ,

ai log
ai

f (b)f (a)
ba

ai
a(i)

n
P

yi

f (x) dx

a+b
2

i=1

Gibbs

 P

xy

xp
p

Carleman

j=1

j=1

1i1 <<ik n

Pr Ai1 Aik

where Ai are events.

Bhatia-Davis
Samuelson

Markov

Chebyshev

2nd

kth

moment

moment

Chernoff



Var[X] M E[X] E[X] m

where X [m, M ].

Paley-Zygmund

n 1 xi + n 1 for i = 1, . . . , n.
P
P
Where =
xi /n , 2 =
(xi )2 /n.

VysochanskijPetunin-Gauss


k




E (X )
and
Pr X t
tk

k/2



nk
for Xi [0, 1] k-wise indep. r.v.,
Pr X t Ck
t2

P
X=
Xi , i = 1, . . . , n, = E[X], Ck = 2 ke1/6k 1.0004, k even.
F (a)/at

Pr[X t]
for X r.v., Pr[X = k] = pk ,
P
k
F (z) = k pk z probability gen. func., and a 1.






e
2
Pr X (1 + )

exp
min{2 + , 3}
(1 + )(1+)
P
where Xi indep. r.v. drawn from {0,1}, X =
Xi , = E[X], 0.






e
2
Pr X (1 )
for [0, 1).

exp
2
(1 )(1)


Simpler (weaker) form: Pr X R 2R for R 2e ( 5.44).


Pr X t

 
n
k

pk
 
t
k



Pr X (1 + )

for Xi {0, 1} k-wise i.r.v., E[Xi ] = p, X =

Hoeffding

Xi [ai , bi ] (w. prob. 1), X =

Kolmogorov

Pn

2 2

i=1

(bi ai )2

Etemadi

Bennett

Bernstein

Azuma

Efron-Stein

McDiarmid

Janson

8
,
3

,
3

if

.
3

2 = Var[X] < , 2 = Var[X] + (E[X] m)2 = E[(X m)2 ].






Pr max |Sk | 3 3 max Pr |Sk |
1kn
1kn
P
where Xi are i.r.v., Sk = ki=1 Xi , 0.

Lov
asz

for martingale (Xk ) and > 0.



n 2
M 
exp 2
where Xi i.r.v.,
M
n 2
i=1
P
1
Var[Xi ], |Xi | M (w. prob. 1), 0,
E[Xi ] = 0, 2 = n
Pr

n
 P

Xi

(u) = (1 + u) log(1 + u) u.


n
 P

2
Pr
Xi exp
2(n 2 + M /3)
i=1

for Xi i.r.v.,

1 P
E[Xi ] = 0, |Xi | < M (w. prob. 1) for all i, 2 = n
Var[Xi ], 0.





2
Pn
for martingale (Xk ) s.t.
Pr Xn X0 2 exp
2 i=1 ci 2


Xi Xi1 < ci (w. prob. 1), for i = 1, . . . , n, 0.

Var[Z]

1
2

n
P

i=1

Z Z (i)

2

for Xi , Xi X i.r.v.,

f : X n R, Z = f (X1 , . . . , Xn ), Z (i) = f (X1 , . . . , Xi , . . . , Xn ).







2 2
for Xi , Xi X i.r.v.,
Pr Z E[Z] 2 exp Pn
2
i=1 ci


Z, Z (i) as before, s.t. Z Z (i) ci for all i, and 0.

V 
M Pr
B i M exp
Q
M = (1 Pr[Bi ]), =

2 2
P

i6=j,Bi Bj

Xi , 0.

Pr

V

 Q
B i (1 xi ) > 0

where Pr[Bi ] for all i,

Pr[Bi Bj ].

where Pr[Bi ] xi

(i,j)D

(1 xj ),

for xi [0, 1) for all i = 1, . . . , n and D the dependency graph.

If each Bi mutually indep. of the set of all other events, exc. at most d,
V 
Pr[Bi ] p for all i = 1, . . . , n , then if ep(d + 1) 1 then Pr
B i > 0.

where X1 , . . . , Xn are i.r.v., E[Xi ] = 0,


P
Var[Xi ] < for all i, Sk = ki=1 Xi and > 0.

if





Pr max1kn |Xk | E |Xn | /

Doob

for Xi i.r.v.,

A related lemma, assuming E[X] = 0, X [a, b] (w. prob. 1) and R:



 2


(b a)2
E eX exp
8


P
1
Pr max |Sk | 2 Var[Sn ] = 12
Var[Xi ]

if

Where X is a unimodal r.v. with mode m,

Xi .

for Xi [0, 1] k-wise i.r.v.,


P

, > 0.
k k = /(1 p), E[Xi ] = pi , X =
Xi , = E[X], p = n

4
92




Pr X E[X]




2
Pr X m 4
92



Pr X m 1

(1+)
n k
p /

k
k




Pr X E[X] 2 exp

for X 0,

Var[X] < , and (0, 1).





Pr |X| a E |X| /a where X is a random variable (r.v.), a > 0.




Pr X c (1 E[X])/(1 c) for X [0, 1] and c 0, E[X] .

Pr X S] E[f (X)]/s for f 0, and f (x) s > 0 for all x S.



Pr X E[X] t Var[X]/t2 where t > 0.


Pr X E[X] t Var[X]/(Var[X] + t2 ) where t > 0.


Pr X > 0 (E[X])2 /(E[X 2 ]) where E[X] 0.


Pr X = 0 Var[X]/(E[X 2 ]) where E[X 2 ] 6= 0.


Var[X]
(1 )2 (E[X])2 + Var[X]



Pr X E[X] 1

L
aszl
o Kozma latest version: http://www.Lkozma.net/inequalities_cheat_sheet

Vous aimerez peut-être aussi