Sever S. Dragomir-Advances in Inequalities of The Schwarz

Advances in Inequalities of the
Schwarz, Triangle and Heisenberg

Type in Inner Product Spaces
Sever Silvestru Dragomir
School of Computer Science & Mathematics, Victoria
University, Melbourne, Victoria, Australia
E-mail address: sever.dragomir@vu.edu.au
URL: http://rgmia.vu.edu.au/SSDragomirWeb.html
2000 Mathematics Subject Classification. Primary 46C05, 46E30;

Secondary 25D15, 26D10
Abstract. The purpose of this book is to give a comprehensive
introduction to several inequalities in Inner Product Spaces that
have important applications in various topics of Contemporary
Mathematics such as: Linear Operators Theory, Partial Differential Equations, Nonlinear Analysis, Approximation Theory, Optimization Theory, Numerical Analysis, Probability Theory, Statistics and other fields.
Contents
Preface
Chapter 1. Inequalities for Hermitian Forms

1.1. Introduction
1.2. Hermitian Forms, Fundamental Properties
1.3. Superadditivity and Monotonicity
1.4. Applications for General Inner Product Spaces
1.5. Applications for Sequences of Vectors
1
1
2
8
15
27
Bibliography
35
Chapter 2. Schwarz Related Inequalities

2.1. Introduction
2.2. Inequalities Related to Schwarzs One
2.3. Kurepa Type Refinements for the Schwarz Inequality
2.4. Refinements of Buzanos and Kurepas Inequalities
2.5. Inequalities for Orthornormal Families
2.6. Generalizations of Precupanu s Inequality
2.7. Some New Refinements of the Schwarz Inequality
2.8. More Schwarz Related Inequalities
37
37
38
46
51
58
65
74
88
Bibliography
105
Chapter 3. Reverses for the Triangle Inequality

3.1. Introduction
3.2. Some Inequalities of Diaz-Metcalf Type
3.3. Additive Reverses for the Triangle Inequality
3.4. Further Additive Reverses
3.5. Reverses of Schwarz Inequality
3.6. Quadratic Reverses of the Triangle Inequality
3.7. Further Quadratic Refinements
3.8. Reverses for Complex Spaces
3.9. Applications for Vector-Valued Integral Inequalities
3.10. Applications for Complex Numbers
107
107
108
112
116
121
122
129
135
142
145
Bibliography
149
iii
iv
CONTENTS
Chapter 4. Reverses for the Continuous Triangle Inequality

4.1. Introduction
4.2. Multiplicative Reverses
4.3. Some Additive Reverses
4.5. Refinements for Complex Spaces
4.6. Applications for Complex-Valued Functions
151
151
152
160
172
178
186
Bibliography
195
Chapter 5. Reverses of the CBS and Heisenberg Inequalities

5.1. Introduction
5.2. Some Reverse Inequalities
5.3. Other Reverses
197
197
198
213
Bibliography
223
Chapter 6. Other Inequalities in Inner Product Spaces

6.1. Bounds for the Distance to Finite-Dimensional Subspaces
6.2. Reversing the CBS Inequality for Sequences
6.3. Other Reverses of the CBS Inequality
225
225
239
259
Bibliography
273
Index
275
Preface
The purpose of this book, that can be seen as a continuation of the
previous one entitled Advances on Inequalities of the Schwarz, Gr
uss
and Bessel Type in Inner Product Spaces (Nova Science Publishers,
NY, 2005), is to give a comprehensive introduction to other classes of
inequalities in Inner Product Spaces that have important applications
in various topics of Contemporary Mathematics such as: Linear Operators Theory, Partial Differential Equations, Nonlinear Analysis, Approximation Theory, Optimization Theory, Numerical Analysis, Probability Theory, Statistics and other fields.
The monograph is intended for use by both researchers in various
fields of Mathematical Inequalities, domains which have grown exponentially in the last decade, as well as by postgraduate students and
scientists applying inequalities in their specific areas.
The aim of Chapter 1 is to present some fundamental analytic properties concerning Hermitian forms defined on real or complex linear
spaces. The basic inequalities as well as various properties of superadditivity and monotonicity for the diverse functionals that can be naturally associated with the quantities involved in the Schwarz inequality
are given. Applications for orthonormal families, Gram determinants,
linear operators defined on Hilbert spaces and sequences of vectors are
also pointed out.
In Chapter 2, classical and recent refinements and reverse inequalities for the Schwarz and the triangle inequalities are presented. Further on, the inequalities obtained by Buzano, Richards, Precupanu
and Moore and their extensions and generalizations for orthonormal
families of vectors in both real and complex inner product spaces are
outlined. Recent results concerning the classical refinement of Schwarz
inequality due to Kurepa for the complexification of real inner product
spaces are also reviewed. Various applications for integral inequalities
including a version of Heisenberg inequality for vector valued functions
in Hilbert spaces are provided as well.
The aim of Chapter 3 is to survey various recent reverses for the
generalised triangle inequality in both its simple form, that are closely
v
vi
PREFACE
related to the Diaz-Metcalf results, or in the equivalent quadratic form

that maybe be of interest in the Geometry of Inner product Spaces.
Applications for vector valued integral inequalities and for complex
numbers are given as well.
Further on, in Chapter 4, some recent reverses of the continuous triangle inequality for Bochner integrable functions with values in Hilbert
spaces and defined on a compact interval [a, b] R are surveyed. Applications for Lebesgue integrable complex-valued functions that generalise and extend the classical result of Karamata are provided as well.
In Chapter 5 some reverses of the Cauchy-Buniakovsky-Schwarz
vector-valued integral inequalities under various assumptions of boundedness for the functions involved are given. Natural applications for the
Heisenberg inequality for vector-valued functions in Hilbert spaces are
also provided.
The last chapter, Chapter 6, is a potpourri of other inequalities
in inner product spaces. The aim of the first section is to point out
some upper bounds for the distance d (x, M ) from a vector x to a finite
dimensional subspace M in terms of the linearly independent vectors
{x1 , . . . , xn } that span M . As a by-product of this endeavour, some
refinements of the generalisations for Bessels inequality due to several
authors including: Boas, Bellman and Bombieri are obtained. Refinements for the well known Hadamards inequality for Gram determinants are also derived.
In the second and third sections of this last chapter, several reverses
for the Cauchy-Bunyakovsky-Schwarz (CBS) inequality for sequences
of vectors in Hilbert spaces are obtained. Applications for bounding
the distance to a finite-dimensional subspace and in reversing the generalised triangle inequality are also given.
For the sake of completeness, all the results presented are completely proved and the original references where they have been firstly
obtained are mentioned. The chapters are relatively independent and
can be read separately.
The Author,
March, 2005.
CHAPTER 1
Inequalities for Hermitian Forms

1.1. Introduction
Let K be the field of real or complex numbers, i.e., K = R or C and
X be a linear space over K.
Definition 1. A functional (, ) : X X K is said to be a
Hermitian form on X if
(H1) (ax + by, z) = a (x, z) + b (y, z) for a, b K and x, y, z X;
(H2) (x, y) = (y, x) for all x, y X.
The functional (, ) is said to be positive semi-definite on a subspace
Y of X if
(H3) (y, y) 0 for every y Y,
and positive definite on Y if it is positive semi-definite on Y and
(H4) (y, y) = 0, y Y implies y = 0.
The functional (, ) is said to be definite on Y provided that either
(, ) or (, ) is positive semi-definite on Y.
When a Hermitian functional (, ) is positive-definite on the whole
space X, then, as usual, we will call it an inner product on X and will
denote it by h, i .
The aim of this chapter is to present some fundamental analytic
properties concerning Hermitian forms defined on real or complex linear
spaces. The basic inequalities as well as various properties of superadditivity and monotonicity for diverse functionals that can be naturally
associated with the quantities involved in the Schwarz inequality are
given. Applications for orthonormal families, Gram determinants, linear operators defined on Hilbert spaces and sequences of vectors are
also pointed out. The results are completely proved and the original
references where they have been firstly obtained are mentioned.
1
1. INEQUALITIES FOR HERMITIAN FORMS
1.2. Hermitian Forms, Fundamental Properties

1.2.1. Schwarzs Inequality. We use the following notations related to a given Hermitian form (, ) on X :
X0 := {x X| (x, x) = 0} ,
K := {x X| (x, x) < 0}
and, for a given z X,
X (z) := {x X| (x, z) = 0}
and L (z) := {az|a K} .
The following fundamental facts concerning Hermitian forms hold

[5]:
Theorem 1 (Kurepa, 1968). Let X and (, ) be as above.
(1) If e X is such that (e, e) 6= 0, then we have the decomposition
M
(1.1)
X = L (e)
X (e) ,
L
where
denotes the direct sum of the linear subspaces X (e)
and L (e) ;
(2) If the functional (, ) is positive semi-definite on X (e) for at
least one e K, then (, ) is positive semi-definite on X (f ) for
each f K;
(3) The functional (, ) is positive semi-definite on X (e) with e
K if and only if the inequality
(1.2)
|(x, y)|2 (x, x) (y, y)
holds for all x K and all y X;

(4) The functional (, ) is semi-definite on X if and only if the
Schwarzs inequality
(1.3)
|(x, y)|2 (x, x) (y, y)
holds for all x, y X;

(5) The case of equality holds in (1.3) for x, y X and in (1.2),
for x K, y X, respectively; if and only if there exists a
scalar a K such that
(x)
y ax X0 := X0 X (x) .
Proof. We follow the argument in [5].
If (e, e) 6= 0, then the element
x := y
(y, e)
e
(e, e)
1.2. HERMITIAN FORMS, FUNDAMENTAL PROPERTIES
has the property that (x, e) = 0, i.e., x X (e) . This proves that X is
a sum of the subspaces L (e) and X (s) . The fact that the sum is direct
is obvious.
Suppose that (e, e) 6= 0 and that (, ) is positive semi-definite on
X. Then for each y X we have y = ae + z with a K and z X (e) ,
from where we get
|(e, y)|2 (e, e) (y, y) = (e, e) (z, z) .
(1.4)
From (1.4) we get the inequality (1.3), with x = e, in the case that
(e, e) > 0 and (1.2) in the case that (e, e) < 0. In addition to this, from
(1.4) we observe that the case of equality holds in (1.2) or in (1.3) if
(e)
and only if (z, z) = 0, i.e., if and only if y ae X0 .
Conversely, if (1.3) holds for all x, y X, then (x, x) has the same
sign over the whole of X, i.e., (, ) is semi-definite on X. In the same
manner, from (1.2), for y X (e) , we get (e, e) (y, y) 0, which implies
(y, y) 0, i.e., (, ) is positive semi-definite on X (e) .
Now, suppose that (, ) is positive semi-definite on X (e) for at least
one e K. Let us prove that (, ) is positive semi-definite on X (f ) for
each f K.
For a given f K, consider the vector
(e, f )
(1.5)
e0 := e
f.
(f, f )
Now,
(e, e) (f, f ) |(e, f )|2
,
(e , e ) = (e , e) =
(f, f )
and together with
0
|(e, y)|2 (e, e) (y, y)
(e0 , f ) = 0
for any y X
imply (e0 , e0 ) 0.
There are two cases to be considered: (e0 , e0 ) > 0 and (e0 , e0 ) = 0.
If (e0 , e0 ) > 0, then for any x X (f ) , the vector
x0 := x ae0
with a =
(x, e0 )
(e0 , e0 )
satisfies the conditions

(x0 , e) = 0 and
(x0 , f ) = 0
which implies
x0 X (e)
and
(x, x) = |a|2 (e0 , e0 ) + (x0 , x0 ) 0.
Therefore (, ) is a positive semi-definite functional on X (f ) .
From the parallelogram identity:

(1.6) (x + y, x + y) + (x y, x y) = 2 [(x, x) + (y, y)] ,
x, y X
(e)
we conclude that the set X0 = X0 X (e) is a linear subspace of X.

Since
1
(1.7)
(x, y) = [(x + y, x + y) + (x y, x y)] , x, y X
4
in the case of real spaces, and
1
[(x + y, x + y) + (x y, x y)]
4
i
+ [(x + iy, x + iy) (x iy, x iy)] , x, y X
4
in the case of complex spaces, hence (x, y) = 0 provided that x and y
(e)
belong to X0 .
If (e0 , e0 ) = 0, then (e0 , e) = (e0 , e0 ) = 0 and then we can conclude
(e)
that e0 X0 . Also, since (e0 , e0 ) = 0 implies (e, f ) 6= 0, hence we have
(1.8) (x, y) =
f = b (e e0 )
with b =
(f, f )
.
(e, f )
Now write
(e)
X (e) = X0
(e)
(e)
X+ ,
(e)
where X+ is any direct complement of X0 in the space X (e) . If y 6= 0,

(e)
then y X+ implies (y, y) > 0. For such a vector y, the vector
y 0 := e0
(e0 , y)
y.
(y, y)
is in X (e) and therefore (y 0 , y 0 ) 0.

On the other hand
|(e0 , y)|2
(y , y ) = (e , y ) =
.
(y, y)
0
(e)
Hence y X+ implies that (e0 , y) = 0, i.e.,

(e, y) =
(e, f )
(f, y) ,
(f, f )
(e)
which together with y X (e) leads to (f, y) = 0. Thus y X+ implies

y X (f ) .
(e)
On the other hand x X0 and f = b (e e0 ) imply (f, x) =
(e)
b (e0 , x) = 0 due to the fact that e0 , x X0 .
(e)
Hence x X0 implies (x, f ) = 0, i.e., x X (f ) .

(e)
(e)
From X0 X (f ) and
X+ X (f ) we get X (e) X (f ) . Since
L
e
/ X (f ) and X = L (e) X (e) , we deduce X (e) = X (f ) and then (, )
is positive semi-definite on X (f ) .
The theorem is completely proved.
In the case of complex linear spaces we may state the following
result as well [5]:
Theorem 2 (Kurepa, 1968). Let X be a complex linear space and
(, ) a hermitian functional on X.
(1) The functional (, ) is semi-definite on X if and only if there
exists at least one vector e X with (e, e) 6= 0 such that
(1.9)
[Re (e, y)]2 (e, e) (y, y) ,
for all y X;
(2) There is no nonzero Hermitian functional (, ) such that the
inequality
(1.10)
[Re (e, y)]2 (e, e) (y, y) ,
(e, e) 6= 0,
holds for all y X and for an e X.

Proof. We follow the proof in [5].
Let and be real numbers and x X (e) a given vector. For
y := ( + i ) e + x we get
(1.11)
[Re (e, y)]2 (e, e) (y, y) = 2 (e, e)2 (e, e) (x, x) .
If (, ) is semi-definite on X, then (1.11) implies (1.9).

Conversely, if (1.9) holds for all y X and for at least one e X,
then (, ) is semi-definite on X (e) . But (1.9) and (1.11) for = 0 lead
to (e, e) (x, x) 0 from which it follows that (e, e) and (x, x) are of
the same sign so that (, ) is semi-definite on X.
Suppose that (, ) 6= 0 and that (1.10) holds. We can assume that
(e, e) < 0. Then (1.10) implies that (, ) is positive semi-definite on
X (e) . On the other hand, if is such that
2 >
(x, x)
,
(e, e)
then (1.11) leads to [Re (e, y)]2 < (e, e) (y, y), contradicting (1.10).
Hence, if a Hermitian functional (, ) is not semi-definite and if
(e, e) 6= 0, then the function y 7 [Re (e, y)]2 (e, e) (y, y) takes
both positive and negative values.
The theorem is completely proved.
1.2.2. Schwarzs Inequality for the Complexification of a

Real Space. Let X be a real linear space. The complexification XC
of X is defined as a complex linear space X X of all ordered pairs
{x, y} (x, y X) endowed with the operations:
{x, y} + {x0 , y 0 } := {x + x0 , y + y 0 } ,
( + i ) {x, y} := {x y, x + y} ,
where x, y, x0 , y 0 X and , R (see for instance [6]).
If z = {x, y} , then we can define the conjugate vector z of z by
z := {x, y} . Similarly, with the scalar case, we denote
Re z = {x, 0}
and
Im z := {0, y} .
Formally, we can write z = x + iy = Re z + i Im z and z = x iy =

Re z i Im z.
Now, let (, ) be a Hermitian functional on X. We may define on
the complexification XC of X, the complexification of (, ) , denoted by
(, )C and defined by:
(x + iy, x0 + iy 0 )C := (x, x0 ) + (y, y 0 ) + i [(y, x0 ) (x, y 0 )] ,
for x, y, x0 , y 0 X.
The following result may be stated [5]:
Theorem 3 (Kurepa, 1968). Let X, XC , (, ) and (, )C be as
above. An inequality of type (1.2) and (1.3) holds for the functional
(, )C in the space XC if and only if the same type of inequality holds
for the functional (, ) in the space X.
Firstly, observe that (, ) is semi-definite if and only if (, )C is
semi-definite.
Now, suppose that e X is such that
|(e, y)|2 (e, e) (y, y) ,
(e, e) < 0
for all y X. Then for x, y X we have

|(e, x + iy)C |2 = [(e, x)]2 + [(e, y)]2
(e, e) [(x, x) + (y, y)]
= (e, e) (x + iy, x + iy)C .
Hence, if for the functional (, ) on X an inequality of type (1.2) holds,
then the same type of inequality holds in XC for the corresponding
functional (, )C .
Conversely, suppose that e, f X are such that
(1.12)
|(e + if, x + iy)C |2 (e + if, e + if )C (x + iy, x + iy)C
holds for all x, y X and that

(1.13)
(e + if, e + if )C = (e, e) + (f, f ) < 0.
If e = af with a real number a, then (1.13) implies that (f, f ) < 0

and (1.12) for y = 0 leads to
[(f, x)]2 (f, f ) (x, x) ,
for all x X. Hence, in this case, we have an inequality of type (1.2)
for the functional (, ) in X.
Suppose that e and g are linearly independent and by Y = L (e, f )
let us denote the subspace of X consisting of all linear combinations of
e and f. On Y we define a hermitian functional D by setting D (x, y) =
(x, y) for x, y Y. Let DC be the complexification of D. Then (1.12)
implies:
(1.14) |DC (e + if, x + iy)|2
DC (e + if, e + if ) DC (x + iy, x + iy) ,
x, y X
and (1.13) implies

(1.15)
D (e, e) + D (f, f ) < 0.
Further, consider in Y a base consisting of the two vectors {u1 , u2 } on

which D is diagonal, i.e., D satisfies
D (x, y) = 1 x1 y1 + 2 x2 y2 ,
where
x = x 1 u1 + x 2 u2 ,
y = y 1 u1 + y 2 u2 ,
and
1 = D (u1 , u1 ) , 2 = D (u2 , u2 ) .
Since for the functional D we have the relations (1.15) and (1.14), we
conclude that D is not a semi-definite functional on Y. Hence 1 2 < 0,
so we can take 1 < 0 and 2 > 0.
Set
X + := {x| (x, e) = (x, f ) = 0, x X} .
Obviously, (x, e) = (x, f ) = 0 if and only if (x1 u1 ) = (x2 u2 ) = 0.
Now, if y X, then the vector
(1.16)
x := y
(y, u1 )
(y, u2 )
u1
u2
(u1 , u1 )
(u2 , u2 )
belongs to X + . From this it follows that

M
X = L (e, f )
X +.
Now, replacing in (1.12) the vector x + iy with z X + , we get from

(1.13) that
[(e, e) + (f, f )] (z, z) 0,
which, together with (1.13) leads to (z, z) 0. Therefore the functional
(, ) is positive semi-definite on X + .
Now, since any y X is of the form (1.16), hence for y X (u1 ) we
get
[(y, u2 )]2
(y, y) = (x, x) +
,
2
which is a nonnegative number. Thus, (, ) is positive semi-definite on
the space X (u1 ) . Since (u1 , u1 ) < 0 we have [(u1 , y)]2 (u1 , u1 ) (y, y)
for any y X and the theorem is completely proved.
1.3. Superadditivity and Monotonicity
1.3.1. The Convex Cone of Nonnegative Hermitian Forms.
Let X be a linear space over the real or complex number field K and
let us denote by H (X) the class of all positive semi-definite Hermitian
forms on X, or, for simplicity, nonnegative forms on X, i.e., the mapping (, ) : X X K belongs to H (X) if it satisfies the conditions
(i) (x, x) 0 for all x in X;
(ii) (x + y, z) = (x, z) + (y, z) for all x, y X and , K
(iii) (y, x) = (x, y) for all x, y X.
1
If (, ) H (X) , then the functional kk = (, ) 2 is a semi-norm on

X and the following equivalent versions of Schwarzs inequality hold:
(1.17)
kxk2 kyk2 |(x, y)|2
or
kxk kyk |(x, y)|
for any x, y X.
Now, let us observe that H (X) is a convex cone in the linear space
of all mappings defined on X 2 with values in K, i.e.,
(e) (, )1 , (, )2 H (X) implies that (, )1 + (, )2 H (X) ;
(ee) 0 and (, ) H (X) implies that (, ) H (X) .
We can introduce on H (X) the following binary relation [1]:
(1.18) (, )2 (, )1
if and only if
kxk2 kxk1
for all x X.
We observe that the following properties hold:

(b) (, )2 (, )1 for all (, ) H (X) ;
(bb) (, )3 (, )2 and (, )2 (, )1 implies that (, )3 (, )1 ;
(bbb) (, )2 (, )1 and (, )1 (, )2 implies that (, )2 = (, )1 ;
1.3. SUPERADDITIVITY AND MONOTONICITY
i.e., the binary relation defined by (1.18) is an order relation on

H (X) .
While (b) and (bb) are obvious from the definition, we should remark, for (bbb), that if (, )2 (, )1 and (, )1 (, )2 , then obviously
kxk2 = kxk1 for all x X, which implies, by the following well known
identity:
(1.19)
(x, y)k

1
:=
kx + yk2k kx yk2k + i kx + iyk2k kx iyk2k
4
with x, y X and k {1, 2}, that (x, y)2 = (x, y)1 for all x, y X.
1.3.2. The Superadditivity and Monotonicity of Mapping.

Let us consider the following mapping [1]:
: H (X) X 2 R+ ,
((, ) ; x, y) := kxk kyk |(x, y)| ,
which is closely related to Schwarzs inequality (1.17).

The following simple properties of are obvious:
(s) ( (, ) ; x, y) = ((, ) ; x, y) ;
(ss) ((, ) ; y, x) = ((, ) ; x, y) ;
(sss) ((, ) ; x, y) 0 (Schwarzs inequality);
for any 0, (, ) H (X) and x, y X.
The following result concerning the functional properties of as a
function depending on the nonnegative hermitian form (, ) has been
obtained in [1]:
Theorem 4 (Dragomir-Mond, 1994). The mapping satisfies the
following statements:
(i) For every (, )i H (X) (i = 1, 2) one has the inequality
(1.20)
((, )1 + (, )2 ; x, y)
((, )1 ; x, y) + ((, )2 ; x, y)
( 0)
for all x, y X, i.e., the mapping (; x, y) is superadditive

on H (X) ;
(ii) For every (, )i H (X) (i = 1, 2) with (, )2 (, )1 one has
(1.21)
((, )2 ; x, y) ((, )1 ; x, y)
( 0)
for all x, y X, i.e., the mapping (; x, y) is nondecreasing

on H (X) .
10
(i) By the Cauchy-Bunyakovsky-Schwarz inequality for real numbers , we have

1
1
a2 + b2 2 c2 + d2 2 ac + bd; a, b, c, d 0.
Therefore,
((, )1 + (, )2 ; x, y)
1
1
= kxk21 + kxk22 2 kyk21 + kyk22 2 |(x, y)1 + (x, y)2 |
kxk1 kyk1 + kxk2 kyk2 |(x, y)1 | |(x, y)2 |
= ((, )1 ; x, y) + ((, )2 ; x, y) ,
for all (, )i H (X) (i = 1, 2) and x, y X, and the statement
is proved.
(ii) Suppose that (, )2 (, )1 and define (, )2,1 := (, )2 (, )1 .
It is obvious that (, )2,1 is a nonnegative hermitian form and
thus, by the above property one has,

((, )2 ; x, y) (, )2,1 + (, )1 ; x, y

(, )2,1 ; x, y + ((, )1 ; x, y)
from where we get:

((, )2 ; x, y) ((, )1 ; x, y) (, )2,1 ; x, y 0
and the proof of the theorem is completed.
Remark 1. If we consider the related mapping [1]
r ((, ) ; x, y) := kxk kyk Re (x, y) ,
then we can show, as above, that (; x, y) is superadditive and nondecreasing on H (X) .
Moreover, if we introduce another mapping, namely, [1]
: H (X) X 2 R+ ,
((, ) ; x, y) := (kxk + kyk)2 kx + yk2 ,
which is connected with the triangle inequality

(1.22)
kx + yk kxk + kyk
for any x, y X
then we observe that

(1.23)
((, ) ; x, y) = 2 r ((, ) ; x, y)
for all (, ) H (X) and x, y X, therefore (; x, y) is in its turn a

superadditive and nondecreasing functional on H (X) .
11
1.3.3. The Superadditivity and Monotonicity of Mapping.

Now consider another mapping naturally associated to Schwarzs inequality, namely [1]
: H (X) X 2 R+ ,
((, ) ; x, y) := kxk2 kyk2 |(x, y)|2 .
It is obvious that the following properties are valid:

(i) ((, ) ; x, y) 0 (Schwarzs inequality);
(ii) ((, ) ; x, y) = ((, ) ; y, x) ;
(iii) ( (, ) ; x, y) = 2 ((, ) ; x, y)
for all x, y X, 0 and (, ) H (X) .
The following theorem incorporates some further properties of this
functional [1]:
Theorem 5 (Dragomir-Mond, 1994). With the above assumptions,
we have:
(i) If (, )i H (X) (i = 1, 2) , then
(1.24) ((, )1 + (, )2 ; x, y) ((, )1 ; x, y) ((, )2 ; x, y)

2
kxk1 kyk1
det
( 0) ;
kxk2 kyk2
i.e., the mapping (; x, y) is strong superadditive on H (X) .
(ii) If (, )i H (X) (i = 1, 2) , with (, )2 (, )1 , then
(1.25) ((, )2 ; x, y) ((, )1 ; x, y)
#!2
"
kxk1
kyk1
1
1
det
kxk22 kxk21 2 kyk22 kyk21 2
( 0) ;
i.e., the mapping (; x, y) is strong nondecreasing on H (X) .

Proof. (i) For all (, )i H (X) (i = 1, 2) and x, y X we have
(1.26)
((, )1 + (, )2 ; x, y)

= kxk22 kxk21 kyk22 kyk21 |(x, y)2 + (x, y)1 |2
kxk22 kyk22 + kxk21 kyk21 + kxk21 kyk22 + kxk22 kyk21
(|(x, y)2 | + |(x, y)1 |)2
= ((, )2 ; x, y) + ((, )1 ; x, y)
+ kxk21 kyk22 + kxk22 kyk21 2 |(x, y)2 (x, y)1 | .
By Schwarzs inequality we have

(1.27)
|(x, y)2 (x, y)1 | kxk1 kyk1 kxk2 kyk2 ,
12
therefore, by (1.26) and (1.27), we can state that

((, )1 + (, )2 ; x, y) ((, )1 ; x, y) ((, )2 ; x, y)
kxk21 kyk22 + kxk22 kyk21 2 kxk1 kyk1 kxk2 kyk2
= (kxk1 kyk2 kxk2 kyk1 )2
and the inequality (1.24) is proved.
(ii) Suppose that (, )2 (, )1 and, as in Theorem 4, define (, )2,1 :=
(, )2 (, )1 . Then (, )2,1 is a nonnegative hermitian form and by (i)
we have

(, )2,1 ; x, y ((, )1 ; x, y)

= (, )2,1 + (, )1 ; x, y ((, )1 ; x, y)

2

kxk1 kyk1
(, )2,1 ; x, y + det
kxk2,1 kyk2,1

2
kxk1 kyk1
det
.
kxk2,1 kyk2,1
Since kzk2,1 = kzk22 kzk21
proved.
12
for z X, hence the inequality (1.25) is
Remark 2. If we consider the functional r ((, ) ; x, y) := kxk2 kyk2

[Re (x, y)]2 , then we can state similar properties for it. We omit the
details.
1.3.4. Superadditivity and Monotonicity of Mapping.
Consider the functional : H (X) X 2 R defined by [2]
1
(1.28)
((, ) ; x, y) = kxk2 kyk2 |(x, y)|2 2 .
1
It is obvious that ((, ) ; x, y) = [ ((, ) ; x, y)] 2 and thus it is monotonic nondecreasing on H (X) . Before we prove that (; x, y) is also
superadditive, which apparently does not follow from the properties of
pointed out in the subsection above, we need the following simple
lemma:
Lemma 1. If (, ) is a nonnegative Hermitian form on X, x, y X
and kyk =
6 0, then
(1.29)
kxk2 kyk2 |(x, y)|2

.
inf kx yk =
K
kyk2
2
13
Proof. Observe that

kx yk2 = kxk2 2 Re [ (x, y)] + ||2 kyk2
and, for kyk =
6 0,

2
kxk2 kyk2 |(x, y)|2 + kyk2 (x, y)
kyk2
h
i
= kxk2 2 Re (x, y) + ||2 kyk2 ,
h
i

(x, y) = Re
(x, y) = Re (x, y) , we deduce the
and since Re
equality

2
2
2
kyk2 (x, y)2
kxk
kyk
|(x,
y)|
+
(1.30)
kx yk2 =
,
kyk2
for any x, y X with kyk =
6 0.
Taking the infimum over K in (1.30), we deduce the desired
result (1.29).
For the subclass J P (X) , of all inner products defined on X, of
H (X) and y 6= 0, we may define
kxk2 kyk2 |(x, y)|2
kyk2
((, ) ; x, y)
=
.
kyk2
((, ) ; x, y) =
The following result may be stated (see also [2]):

Theorem 6 (Dragomir-Mond, 1996). The functional (; x, y) is
superadditive and monotonic nondecreasing on J P (X) for any x, y
X with y 6= 0.
Proof. Let (, )1 , (, )2 J P (X) . Then
(1.31)
((, )1 + (, )2 ; x, y)

kxk21 + kxk22 kyk21 + kyk22 |(x, y)1 + (x, y)2 |2
=
kyk21 kyk22

= inf kx yk21 + kx yk22 ,
K
and for the last equality we have used Lemma 1.
14
Also,
(1.32)
((, )i ; x, y) =
kxk2i kyk2i |(x, y)i |2

kyk2i
= inf kx yk2i ,
K
i = 1, 2.
Utilising the infimum property that

inf (f () + g ()) inf f () + inf g () ,
we can write that

inf kx yk21 + kx yk22 inf kx yk21 + inf kx yk22 ,
K
which proves the superadditivity of (; x, y) .

The monotonicity follows by the superadditivity property and the
theorem is completely proved.
Corollary 1. If (, )i J P (X) with (, )2 (, )1 and x, y X
are such that x, y 6= 0, then:
(
)
kyk22 kxk22
((, )2 ; x, y) max
(1.33)
,
((, )1 ; x, y)
kyk21 kxk21
( ((, )1 ; x, y))
or equivalently, [2]
(1.34) ((, )2 ; x, y) ((, )1 ; x, y)
(
)
kyk22 kyk21 kxk22 kxk21
max
,
((, )1 ; x, y) .
kyk21
kxk21
The following strong superadditivity property of (; x, y) that is
different from the one in Subsection 1.3.2 holds [2]:
Corollary 2 (Dragomir-Mond, 1996). If (, )i J P (X) and
x, y X with x, y 6= 0, then
(1.35) ((, )1 + (, )2 ; x, y) ((, )1 ; x, y) ((, )2 ; x, y)
(
2

2
kyk2
kyk1
((, )1 ; x, y) +
((, )2 ; x, y) ;
max
kyk1
kyk2
)

2

2
kxk2
kxk1
((, )1 ; x, y) +
((, )2 ; x, y)
( 0) .
kxk1
kxk2
1.4. APPLICATIONS FOR GENERAL INNER PRODUCT SPACES
15
Proof. Utilising the identities (1.31) and (1.32) and taking into
account that (; x, y) is superadditive, we can state that
(1.36) ((, )1 + (, )2 ; x, y)
kyk21 + kyk22
kyk21 + kyk22
((, )1 ; x, y) +
((, )2 ; x, y)
kyk21
kyk22
= ((, )1 ; x, y) + ((, )2 ; x, y)

2

2
kyk2
kyk1
+
((, )1 ; x, y) +
((, )2 ; x, y)
kyk1
kyk2
and a similar inequality with x instead of y. These show that the desired
inequality (1.35) holds true.
Remark 3. Obviously, all the inequalities above remain true if
(, )i , i = 1, 2 are nonnegative Hermitian forms for which we have
kxki , kyki 6= 0.
Finally, we may state and prove the superadditivity result for the
mapping (see [2]):
Theorem 7 (Dragomir-Mond, 1996). The mapping defined by
(1.28) is superadditive on H (X) .
Proof. Without loss of generality, if (, )i H (X) and x, y X,
we may assume, for instance, that kyki 6= 0, i = 1, 2.
If so, then

2

2
kyk1
kyk2
((, )1 ; x, y) +
((, )2 ; x, y)
kyk1
kyk2
1
2 [ ((, )1 ; x, y) ((, )2 ; x, y)] 2 ,

and by making use of (1.36) we get:
n
o2
1
1
((, )1 + (, )2 ; x, y) [ ((, )1 ; x, y)] 2 + [ ((, )2 ; x, y)] 2 ,
which is exactly the superadditivity property for .
1.4. Applications for General Inner Product Spaces
1.4.1. Inequalities for Orthonormal Families. Let (H; h, i)
be an inner product space over the real or complex number field K. The
family of vectors E := {ei }iI (I is a finite or infinite) is an orthonormal
family of vectors if hei , ej i = ij for i, j I, where ij is Kroneckers
delta.
16
The following inequality is well known in the literature as Bessels

inequality:
X
(1.37)
|hx, ei i|2 kxk2
iF
for any F a finite part of I and x a vector in H.

If by F (I) we denote the family of all finite parts of I (including
the empty set ), then for any F F (I) \ {} the functional (, )F :
H H K given by
X
(1.38)
(x, y)F :=
hx, ei i hei , yi
iF
is a Hermitian form on H.
It is obvious that if F1 , F2 F (I) \ {} and F1 F2 = , then
(, )F1 F2 = (, )F1 + (, )F2 .
We can define the functional : F (I) H 2 R+ by
(1.39)
(F ; x, y) := kxkF kykF |(x, y)F | ,
where
! 12
kxkF :=
|hx, ei i|2
= [(x, x)F ] 2 ,
x H.
iF
The following proposition may be stated (see also [2]):

Proposition 1 (Dragomir-Mond, 1995). The mapping satisfies
the following
(i) If F1 , F2 F (I) \ {} with F1 F2 = , then
(F1 F2 ; x, y) (F1 ; x, y) + (F2 ; x, y)
( 0)
for any x, y H, i.e., the mapping (; x, y) is an index set

superadditive mapping on F (I) ;
(ii) If 6= F1 F2 , F1 , F2 F (I) , then
(F2 ; x, y) (F1 ; x, y)
( 0) ,
i.e., the mapping (; x, y) is an index set monotonic mapping

on F (I) .
The proof is obvious by Theorem 4 and we omit the details.
We can also define the mapping r (; , ) : F (I) H 2 R+ by
r (F ; x, y) := kxkF kykF Re (x, y)F ,
which also has the properties (i) and (ii) of Proposition 1.
17
Since, by Bessels inequality the hermitian form (, )F h, i in the

sense of definition (1.18) then by Theorem 4 we may state the following
refinements of Schwarzs inequality [1]:
Proposition 2 (Dragomir-Mond, 1994). For any F F (I) \ {0} ,
we have the inequalities
(1.40) kxk kyk |hx, yi|
! 12
! 12
|hx, ei i|
|hy, ei i|
iF
iF

X

hx, ei i hei , yi

iF
and
(1.41)
kxk kyk |hx, yi|

! 21
kxk
X
iF
|hx, ei i|
! 12
2
kyk
|hy, ei i|
iF

X

hx, yi
hx, ei i hei , yi

iF
and the corresponding versions on replacing || by Re () , where x, y are

vectors in H.
Remark 4. Note that the inequality (1.40) and its version for Re ()
has been established for the first time and utilising a different argument
by Dragomir and Sandor in 1994 (see [3, Theorem 5 and Remark 2]).
If we now define the mapping : F (I) H 2 R+ by
(F ; x, y) := kxk2F kyk2F |(x, y)F |2
and making use of Theorem 5, we may state the following result [2].
Proposition 3 (Dragomir-Mond, 1995). The mapping satisfies
the following properties:
(i) If F1 , F2 F (I) \ {} with F1 F2 = , then
(1.42) (F1 F2 ; x, y) (F1 ; x, y) (F2 ; x, y)

2
kxkF1 kykF1
det
kxkF2 kykF2
( 0) ,
i.e., the mapping (; x, y) is strong superadditive as an index

set mapping;
18
(ii) If 6= F1 F2 , F1 , F2 F (I) , then

(1.43) (F2 ; x, y) (F1 ; x, y)
"
kxkF1
1
det
kxk2F2 kxk2F1 2
#!2
kykF1
kyk2F2 kyk2F1
( 0) ,
12
i.e., the mapping (; x, y) is strong nondecreasing as an index

set mapping.
On applying the same general result in Theorem 5, (ii) for the
hermitian functionals (, )F (F F (I) \ {}) and h, i for which, by
Bessels inequality we know that (, )F h, i , we may state the following result as well, which provides refinements for the Schwarz inequality.
Proposition 4 (Dragomir-Mond, 1994). For any F F (I) \ {},
we have the inequalities:
(1.44) kxk2 kyk2 |hx, yi|2
X
iF

2
X

X

2
2
|hx, ei i|
|hy, ei i|
hx, ei i hei , yi

iF
( 0)
iF
and
(1.45) kxk2 kyk2 |hx, yi|2
!
kxk2
|hx, ei i|2
!
kyk2
iF
|hy, ei i|2
iF

2

X

hx, ei i hei , yi
hx, yi

( 0) ,
iF
for any x, y H.
On utilising Corollary 2 we may state the following different superadditivity property for the mapping (; x, y) .
19
Proposition 5. If F1 , F2 F (I) \ {} with F1 F2 = , then

(1.46) (F1 F2 ; x, y) (F1 ; x, y) (F2 ; x, y)
(

kykF2 2
kykF1 2
max
(F1 ; x, y) +
(F2 ; x, y) ;
kykF1
kykF2
)

kxkF2 2
kxkF1 2
(F1 ; x, y) +
(F2 ; x, y)
( 0)
kxkF1
kxkF2
for any x, y H\ {0} .
Further, for y
/ M where M = Sp {ei }iI is the linear space
spanned by E = {ei }iI , we can also consider the functional : F (I)
H 2 R+ defined by
kxk2F kyk2F |(x, y)F |2
(F ; x, y)
(F ; x, y) :=
=
,
kyk2F
kyk2F
where x H and F 6= .
Utilising Theorem 6, we may state the following result concerning
the properties of the functional (; x, y) with x and y as above.
Proposition 6. For any x H and y H\M , the functional
(; x, y) is superadditive and monotonic nondecreasing as an index set
mapping on F (I) .
Since h, i (, )F for any F F (I) , on making use of Corollary
1, we may state the following refinement of Schwarzs inequality:
Proposition 7. Let x H and y H\MF , where MF := Sp {ei }iI
and F F (I) \ {} is given. Then
(
)
2
2
kyk
kxk
(1.47) kxk2 kyk2 |hx, yi|2 max P
2, P
2
iF |hy, ei i|
iF |hx, ei i|

2
X

X
X

2
2
|hx, ei i|
|hy, ei i|
hx, ei i hei , yi

iF
iF
iF

2

X
X
X

|hx, ei i|2
|hy, ei i|2
hx, ei i hei , yi ,

iF
iF
iF
which is a refinement of (1.45) in the case that y H\MF .

Finally, consider the functional : F (I) H 2 R+ given by
1
1
(F ; x, y) := [ (F ; x, y)] 2 = kxk2F kyk2F |(x, y)F |2 2 .
20
Utilising Theorem 7, we may state the following:

Proposition 8. The functional (; x, y) is superadditive as an
index set mapping on F (I) for each x, y H.
As a dual approach, one may also consider the following form
(, )C,F : H H R given by:
(1.48)
(x, y)C,F := hx, yi (x, y)F = hx, yi
hx, ei i hei , yi .
iF
By Bessels inequality, we observe that (, )C,F is a nonnegative hermitian form and, obviously
(, )I + (, )C,F = h, i .
Utilising the superadditivity properties from Section 1.3, one may state
the following refinement of the Schwarz inequality:
! 21
X
|hx, ei i|2
|hy, ei i|2
iF
iF
! 12
+
kxk2
|hx, ei i|2

X

hx, ei i hei , yi

iF
! 12
X
kyk2
|hy, ei i|2
iF
iF

X

hx, yi
hx, ei i hei , yi

( 0) ,
iF
(1.50) kxk2 kyk2 |hx, yi|2

2
X

X

|hx, ei i|2
|hy, ei i|2
hx, ei i hei , yi

iF
iF
iF
!
!
X
X
+ kxk2
|hx, ei i|2
kyk2
|hy, ei i|2
X
iF
iF

2

X

hx, yi
hx, ei i hei , yi

iF
( 0)
21
and
(1.51)
1
kxk2 kyk2 |hx, yi|2 2

2 12
X

X
X

|hx, ei i|2
|hy, ei i|2
hx, ei i hei , yi

iF
iF
iF
!
!
"
X
X
+
kxk2
|hx, ei i|2
kyk2
|hy, ei i|2
iF
iF

2 12

X

hx, yi
hx, ei i hei , yi

( 0) ,
iF
for any x, y H and F F (I) \ {} .

1.4.2. Inequalities for Gram Determinants. Let {x1 , . . . , xn }
be vectors in the inner product space (H, h, i) over the real or complex
number field K. Consider the gram matrix associated to the above
vectors:
hx1 , x1 i hx1 , x2 i hx1 , xn i

hx2 , x1 i
hx2 , xn i
.
G (x1 , . . . , xn ) :=

hxn , x1 i hxn , x2 i hxn , xn i
The determinant
(x1 , . . . , xn ) := det G (x1 , . . . , xn )
is called the Gram determinant associated to the system {x1 , . . . , xn } .
If {x1 , . . . , xn } does not contain the null vector 0, then [4]
(1.52)
0 (x1 , . . . , xn ) kx1 k2 kx2 k2 kxn k2 .
The equality holds on the left (respectively right) side of (1.52) if

and only if {x1 , . . . , xn } is linearly dependent (respectively orthogonal). The first inequality in (1.52) is known in the literature as Grams
inequality while the second one is known as Hadamards inequality.
The following result obtained in [3] may be regarded as a refinement
of Grams inequality:
Theorem 8 (Dragomir-Sandor, 1994). Let {x1 , . . . , xn } be a system
of nonzero vectors in H. Then for any x, y H one has:
(1.53)
(x, x1 , . . . , xn ) (y, x1 , . . . , xn ) | (x1 , . . . , xn ) (x, y)|2 ,
22
where (x1 , . . . , xn ) (x, y) is defined by:

(x1 , . . . , xn ) (x, y)
hx, yi hx, x1 i
hx, xn i
hx1 , yi
.
:= det

G (x1 , . . . , xn )
hxn , yi
Proof. We follow the proof from [3].

Let us consider the mapping p : H H K given by
p (x, y) = (x1 , . . . , xn ) (x, y) .
Utilising the properties of determinants, we notice that
p (x, y) = (x, x1 , . . . , xn ) 0,
p (x + y, z) = (x1 , . . . , xn ) (x + y, z)
= (x1 , . . . , xn ) (x, z) + (x1 , . . . , xn ) (y, z)
= p (x, z) + p (y, z) ,
p (x, y) = p (x, y) ,
p (y, x) = p (x, y),
for any x, y, z H and K, showing that p (, ) is a nonnegative
hermitian from on X. Writing Schwarzs inequality for p (, ) we deduce
the desired result (1.53).
In a similar manner, if we define q : H H K by
q (x, y) := (x, y)
n
Y
kxi k2 p (x, y)
i=1
= (x, y)
n
Y
kxi k2 (x1 , . . . , xn ) (x, y) ,
i=1
then, using Hadamards inequality, we conclude that q (, ) is also a

nonnegative hermitian form. Therefore, by Schwarzs inequality applied for q (, ) , we can state the following result as well [3]:
23
Theorem 9 (Dragomir-Sandor, 1994). With the assumptions of

Theorem 8, we have:
"
#
n
Y
2
2
(1.54)
kxk
kxi k (x, x1 , . . . , xn )
i=1
"
kyk
n
Y
#
2
kxi k (y, x1 , . . . , xn )
i=1

2
n

Y

hx, yi
kxi k2 (x1 , . . . , xn ) (x, y) ,

i=1
for each x, y H.
Observing that, for a given set of nonzero vectors {x1 , . . . , xn } ,
n
Y
p (x, y) + q (x, y) = (x, y)
kxi k2 ,
i=1
for any x, y H, then, on making use of the superadditivity properties of the various functionals defined in Section 1.3, we can state the
following refinements of the Schwarz inequality in inner product spaces:
(1.55) [kxk kyk |hx, yi|]
n
Y
kxi k2
i=1
1
[ (x, x1 , . . . , xn ) (y, x1 , . . . , xn )] 2 | (x1 , . . . , xn ) (x, y)|

"
# 12
n
Y
+ kxk2
kxi k2 (x, x1 , . . . , xn )
i=1
"
kyk
n
Y
2
# 12
kxi k2 (y, x1 , . . . , xn )
i=1

n

Y

2
hx, yi
kxi k (x1 , . . . , xn ) (x, y)

( 0) ,
i=1
n
Y

2
2
2
kxi k4
i=1
(x, x1 , . . . , xn ) (y, x1 , . . . , xn ) | (x1 , . . . , xn ) (x, y)|2

"
#
n
Y
+ kxk2
kxi k2 (x, x1 , . . . , xn )
i=1
24
"
kyk
n
Y
#
2
kxi k (y, x1 , . . . , xn )
i=1

2
n

Y

hx, yi
kxi k2 (x1 , . . . , xn ) (x, y)

( 0) ,
i=1
and
(1.57)
[kxk kyk |hx, yi|]
1
2
n
Y
kxi k2
i=1

1
(x, x1 , . . . , xn ) (y, x1 , . . . , xn ) | (x1 , . . . , xn ) (x, y)|2 2
("
#
n
Y
+
kxk2
kxi k2 (x, x1 , . . . , xn )
"
kyk2
i=1
n
Y
kxi k2 (y, x1 , . . . , xn )
i=1

2 21
n

Y

hx, yi
kxi k2 (x1 , . . . , xn ) (x, y)

( 0) .
i=1
1.4.3. Inequalities for Linear Operators. Let A : H H be

a linear bounded operator and
kAk := sup {kAxk , kxk < 1}
its norm.
If we consider the hermitian forms (, )2 , (, )1 : H H defined
by
(x, y)1 := hAx, Ayi ,
(x, y)2 := kAk2 hx, yi
then obviously (, )2 (, )1 in the sense of definition (1.18) and utilising the monotonicity properties of the functional considered in Section
1.3, we may state the following inequalities:
(1.58) kAk2 [kxk kyk |hx, yi|] kAxk kAyk |hAx, Ayi|
( 0) ,

(1.59) kAk4 kxk2 kyk2 |hx, yi|2
kAxk2 kAyk2 |hAx, Ayi|2
( 0)
for any x, y H, and the corresponding versions on replacing || by

Re () .
The results (1.58) and (1.59) have been obtained by Dragomir and
Mond in [1].
25
On using Corollary 1, we may deduce the following inequality as

well:

(1.60) kAk2 kxk2 kyk2 |hx, yi|2
(
)

kxk2 kyk2
max
kAxk2 kAyk2 |hAx, Ayi|2
( 0)
2,
2
kAxk kAyk
for any x, y H with Ax, Ay 6= 0; which improves (1.59) for x, y
specified before.
Similarly, if B : H H is a linear operator satisfying the condition
(1.61)
kBxk m kxk
for any x H,
where m > 0 is given, then the hermitian forms [x, y]2 := hBx, Byi ,
[x, y]1 := m2 hx, yi , have the property that [, ]2 [, ]1 . Therefore,
from the monotonicity results established in Section 1.3, we can state
that
(1.62) kBxk kByk |hBx, Byi| m2 [kxk kyk |hx, yi|]
(1.63)
kBxk2 kByk2 |hBx, Byi|2

m4 kxk2 kyk2 |hx, yi|2
( 0) ,
( 0)
for any x, y H, and the corresponding results on replacing || by

Re () .
The same Corollary 1, would give the inequality
(1.64) kBxk2 kByk2 |hBx, Byi|2
(
)
kBxk2 kByk2
2
2
2
2
m max
,
kxk
kyk
|hx,
yi|
kxk2 kyk2
for x, y 6= 0, which is an improvement of (1.63).
We recall that a linear self-adjoint operator P : H H is nonnegative if hP x, xi 0 for any x H. P is called positive if hP x, xi = 0
and positive definite with the constant > 0 if hP x, xi kxk2 for
any x H.
If A, B : H H are two linear self-adjoint operators such that
A B (this means that A B is nonnegative), then the corresponding
hermitian forms (x, y)A := hAx, yi and (x, y)B := hBx, yi satisfies the
property that (, )A (, )B .
If by P (H) we denote the cone of all linear self-adjoint and nonnegative operators defined in the Hilbert space H, then, on utilising
26
the results of Section 1.3, we may state that the functionals 0 , 0 , 0 :

P (H) H 2 [0, ] given by
1
0 (P ; x, y) := hAx, xi 2 hP y, yi 2 |hP x, yi| ,

0 (P ; x, y) := hP x, xi hP y, yi |hP x, yi|2 ,

1
0 (P ; x, y) := hP x, xi hP y, yi |hP x, yi|2 2 ,
are superadditive and monotonic decreasing on P (H) , i.e.,
0 (P + Q; x, y) 0 (P ; x, y) + 0 (Q; x, y)
( 0)
for any P, Q P (H) and x, y H, and

0 (P ; x, y) 0 (Q; x, y)
( 0)
for any P, Q with P Q 0 and x, y H, where {, , } .

The superadditivity and monotonicity properties of 0 and 0 have
been noted by Dragomir and Mond in [1].
If u P (H) is such that I U 0, where I is the identity
operator, then on using the superadditivity property of the functionals
0 , 0 and 0 one may state the following refinements for the Schwarz
inequality:
1
(1.65) kxk kyk |hx, yi| hU x, xi 2 hU y, yi 2 |hU x, yi|

1
+ h(I U ) x, xi 2 h(I U ) y, yi 2 |h(I U ) x, yi|
( 0) ,
(1.66) kxk2 kyk2 |hx, yi|2 hU x, xi hU y, yi |hU x, yi|2

+ h(I U ) x, xi h(I U ) y, yi |h(I U ) x, yi|2
( 0) ,
and
(1.67)
1
hU x, xi hU y, yi |hU x, yi|2 2
1
+ h(I U ) x, xi h(I U ) y, yi |h(I U ) x, yi|2 2
( 0)
kxk2 kyk2 |hx, yi|2
12
for any x, y H.
Note that (1.67) is a better result than (1.66).
Finally, if we assume that D P (H) with D I, where >
0, i.e., D is positive definite on H, then we may state the following
inequalities
1
(1.68) hDx, xi 2 hDy, yi 2 |hDx, yi| [kxk kyk |hx, yi|]

(1.69)
hDx, xi hDy, yi |hDx, yi|2

2 kxk2 kyk2 |hx, yi|2
( 0) ,
( 0) ,
1.5. APPLICATIONS FOR SEQUENCES OF VECTORS
27
for any x, y H and

(1.70) hDx, xi hDy, yi |hDx, yi|2

hDx, xi hDy, yi
kxk2 kyk2 |hx, yi|2
max
2 ,
2
kxk
kyk
( 0)
for any x, y H\ {0} .

The results (1.68) and (1.69) have been obtained by Dragomir and
Mond in [1].
Note that (1.70) is a better result than (1.69).
The above results (1.65) (1.70) also hold for Re () instead of || .
1.5. Applications for Sequences of Vectors
1.5.1. The Case of Mapping . Let Pf (N) be the family of
finite parts of the natural number set N, S+ (R) the cone of nonnegative
real sequences and for a given inner product space (H; h, i) over the
real or complex number field K, S (H) the linear space of all sequences
of vectors from H, i.e.,

S (H) := x|x = (xi )iN , xi H, i N .
Consider h, ip,I : S (H) S (H) R defined by
X
pi hxi , yi i .
hx, yip,I :=
iI
We may define the mapping by

! 21
(1.71) (p, I, x, y) :=
X
iI
pi kxi k
X
2
pi kyi k2
iI

X

pi hxi , yi i ,

iI
where p S+ (R) , I Pf (N) and x, y S (H) .

We observe that, for a I Pf (N) \ {} , the functional h, ip,I
h, iq,I , provided p q 0.
Using Theorem 4, we may state the following result.
Proposition 9. Let I Pf (N) \ {}, x, y S (H) . Then the
functional (, I, x, y) is superadditive and monotonic nondecreasing
on S+ (R) .
If I, J Pf (N) \ {} , with I J = , for a given p S+ (R) , we
observe that
(1.72)
h, ip,IJ = h, ip,I + h, ip,J .
Taking into account this property and on making use of Theorem 4,

we may state the following result.
28
Proposition 10. Let p S+ (R) and x, y S (H) .

(i) For any I, J Pf (N) \ {} , with I J = , we have
(1.73)
(p, I J, x, y) (p, I, x, y) + (p, J, x, y)
( 0) ,
i.e., (p, , x, y) is superadditive as an index set mapping on

Pf (N) .
(ii) If 6= J I, I, J Pf (N) , then
(p, I, x, y) (p, J, x, y)
(1.74)
( 0) ,
i.e., (p, , x, y) is monotonic nondecreasing as an index set

mapping on S+ (R) .
It is well known that the following Cauchy-Bunyakovsky-Schwarz
(CBS) type inequality for sequences of vectors in an inner product
space holds true:

2
X

X
X

2
2
(1.75)
pi kxi k
pi kyi k
pi hxi , yi i

iI
iI
iI
for I Pf (N) \ {} , p S+ (R) and x, y S (H) .

If pi > 0 for all i I, then equality holds in (1.75) if and only if
there exists a scalar K such that xi = yi , i I.
Utilising the above results for the functional , we may state the
following inequalities related to the (CBS)-inequality (1.75).
(1) Let i R, xi , yi H, i {1, . . . , n} . Then one has the
inequality:

n
X

(1.76)
kxi k
kyi k
hxi , yi i

i=1
i=1
i=1

! 12 n
n
n

X
X
X

hxi , yi i sin2 i
kxi k2 sin2 i
kyi k2 sin2 i

i=1
i=1
i=1
! 12
n
n
X
X
+
kxi k2 cos2 i
kyi k2 cos2 i
n
X
n
X
i=1
i=1
n

X

hxi , yi i cos2 i 0.

i=1
29
(2) Denote Sn (1) := {p S+ (R) |pi 1 for all i {1, . . . , n}} .

Then for all xi , yi H, i {1, . . . , n} , we have the bound:
! 12

n
X

kxi k
kyi k

hxi , yi i
(1.77)

i=1
i=1
i=1

! 21 n
n
n
X

X
X

2
2

pi hxi , yi i 0.
= sup
pi kxi k
pi kyi k

pSn (1)
n
X
n
X
i=1
i=1
i=1
(3) Let pi 0, xi , yi H, i {1, . . . , n} . Then we have the

inequality:
! 12

2n
X

(1.78)
pi kxi k
pi kyi k

pi hxi , yi i

i=1
i=1
i=1

! 12 n
n
n
X

X
X

p2k kx2k k2
p2k ky2k k2

p2k hx2k , y2k i

k=1
k=1
k=1
! 12
n
n
X
X
+
p2k1 kx2k1 k2
p2k1 ky2k1 k2
2n
X
2n
X
k=1
k=1

n
X

p2k1 hx2k1 , y2k1 i

( 0) .
k=1
(4) We have the bound:

# 12

n
X

(1.79)

pi hxi , yi i

i=1
i=1
i=1
"

# 12
X

X
X

=
sup
pi kxi k2
pi kyi k2
pi hxi , yi i 0.

6=I{1,...,n}
"
n
X
n
X
2
pi kxi k
pi kyi k2
iI
iI
iI
(5) The sequence Sn given by

Sn :=
n
X
i=1
pi kxi k
n
X
2
! 12
pi kyi k2
i=1

n

X

pi hxi , yi i

i=1
is nondecreasing, i.e.,
(1.80)
Sk+1 Sk ,
k2
30
and we have the bound

n
1
1
(1.81) Sn max
pi kxi k2 + pj kxj k2 2 pi kyi k2 + pj kyj k2 2
1i<jn

|pi hxi , yi i + pj hxj , yj i| 0,
for n 2 and xi , yi H, i {1, . . . , n} .
Remark 5. The results in this subsection have been obtained by
Dragomir and Mond in [1] for the particular case of scalar sequences
x and y.
1.5.2. The Case of Mapping . Under the assumptions of the
above subsection, we can define the following functional

2
X

X
X

(p, I, x, y) :=
pi kxi k2
pi kyi k2
pi hxi , yi i ,

iI
iI
iI
where p S+ (R) , I Pf (N) \ {} and x, y S (H) .

Utilising Theorem 5, we may state the following results.
Proposition 11. We have
(i) For any p, q S+ (R) , I Pf (N) \ {} and x, y S (H) we
have
(1.82) (p + q, I, x, y) (p, I, x, y) (q, I, x, y)

21
12
P
P
2
2
pi kxi k
pi kyi k
iI
det iI
12
21
P
P
2
2
qi kyi k
qi kxi k
0.
iI
iI
(ii) If p q 0, then
(1.83) (p, I, x, y) (q, I, x, y)

21

12
P
P
2
2
pi kyi k
pi kxi k
iI
det iI
12
21
P
P
2
2
(pi qi ) kyi k
(pi qi ) kxi k
iI
iI
0.
31
(i) For any I, J Pf (N) , with I J = and p S+ (R) , x, y

S (H) , we have
(1.84) (p, I J, x, y) (p, I, x, y) (p, J, x, y)

21
12
P
P
2
2
pi kxi k
pi kyi k
iI
det iI
12
12
P
P
2
2
pi kxi k
pi kyi k
iJ
0.
iJ
(ii) If 6= J I, I 6= J, I, J Pf (N) , then we have

(1.85) (p, I, x, y) (p, J, x, y)

12

12
P
P
2
2
pi kxi k
pi kyi k
iI
iI
! 21
! 12
det
P
P
2
2
pi kxi k
pi kyi k
iI\J
0.
iI\J
The following particular instances that provide refinements for the

(CBS)-inequality may be stated as well:

2
X

X
X

(1.86)
kxi k2
kyi k2
hxi , yi i

iI
iI
iI

2
X

X
X

kxi k2 sin2 i
kyi k2 sin2 i
hxi , yi i sin2 i

iI
iI
iI
X
X
+
kxi k2 cos2 i
kyi k2 cos2 i
iI
iI

2
X

hxi , yi i cos2 i

iI

21
12
P
P
2
2
2
2
kxi k sin i
kyi k sin i
iI
iI
det
12
12
P
P
2
2
2
2
kyi k cos i
kxi k cos i
iI
iI
0,
where xi , yi H, i R, i I and I Pf (N) \ {} .
32
Suppose that pi 0, xi , yi H, i {1, . . . , 2n} . Then

2
2n
2n
X

X

pi kxi k2
pi kyi k2
pi hxi , yi i
(1.87)

i=1
i=1
i=1

2
n
n
n
X

X
X

2
2
p2k kx2k k
p2k ky2k k
p2k hx2k , y2k i

2n
X
k=1
k=1
n
X
p2k1 kx2k1 k2
k=1
k=1
n
X
p2k1 ky2k1 k2
k=1

2
n
X

p2k1 hx2k1 , y2k1 i

k=1

21
n
12
n
P
P
2
2
p2k kx2k k
p2k ky2k k

k=1
k=1
det
12 n
12
n
P
P
2
2
p2k1 kx2k1 k
p2k1 ky2k1 k
k=1
k=1
0.
Remark 6. The above results (1.82) (1.87) have been obtained
for the case where x and y are real or complex numbers by Dragomir
and Mond [1].
Further, if we use Corollaries 2 and 1, then we can state the following propositions as well.
(i) For any p, q S+ (R) , I Pf (N) \ {} and x, y S (H) \ {0}
we have
(1.88) (p + q, I, x, y) (p, I, x, y) (q, I, x, y)
(P
P
2
2
iI pi kxi k
iI qi kxi k
max P
(q,
I,
x,
y)
+
P
2
2 (p, I, x, y) ,
iI qi kxi k
iI pi kxi k
)
P
P
2
2
p
ky
k
q
ky
k
i
i
i
iI i
0.
PiI
2 (q, I, x, y) + P
2 (p, I, x, y)
q
ky
k
p
ky
k
i
i
i
i
iI
iI
33
(ii) If p q 0 and I Pf (N) \ {}, x, y S (H) \ {0} , then:

(1.89) (p, I, x, y) (q, I, x, y)
(P
)
2 P
2
(p
q
)
ky
k
(p
q
)
kx
k
i
i
i
i
i
i
iI
max
, iI
(p, I, x, y) 0.
P
P
2
2
iI pi kxi k
iI pi kyi k
(i) For any I, J Pf (N) \ {} , with IJ = and p S+ (R) , x, y
S (H) \ {0} , we have
(1.90) (p, I J, x, y) (p, I, x, y) (p, J, x, y)
(P
P
2
2
jJ pj kxj k
iI pi kxi k
max P
2 (p, J, x, y) + P
2 (p, I, x, y) ,
jJ pj kxj k
iI pi kxi k
)
P
P
2
2
jJ pj kyj k
iI pi kyi k
0.
P
2 (p, J, x, y) + P
2 (p, I, x, y)
jJ pj kyj k
iI pi kyi k
(ii) If 6= J I, I 6= J, I, J Pf (N) \ {} and p S+ (R) \ {0} ,
x, y S (H) \ {0} , then
(1.91) (p, I, x, y) (p, J, x, y)
(P
)
2 P
2
kI\J pk kxk k
kI\J pk kyk k
max
(p, J, x, y) 0.
P
2 , P
2
iI pi kxi k
iI pi kyi k
Remark 7. The results in Proposition 13 have been obtained by
Dragomir and Mond in [2] for the case of scalar sequences x and y.
1.5.3. The Case of Mapping . With the assumptions in the
first subsections, we can define the following functional
1
(p, I, x, y) := [ (p, I, x, y)] 2

2 21

X
X
X

=
pi kxi k2
pi kyi k2
pi hxi , yi i ,

iI
iI
iI
where p S+ (R) , I Pf (N) \ {} and x, y S (H) .

Utilising Theorem 7, we can state the following results:
(i) The functional (, I, x, y) is superadditive on S+ (R) for any
I Pf (N) \ {} and x, y S (H) .
(ii) The functional (p, , x, y) is superadditive as an index set
mapping on Pf (N) and x, y S (H) .
34
As simple consequences of the above proposition, we may state the

following refinements of the (CBS)-inequality.
(a) If x, y S (H) and i R, i I with I Pf (N) \ {} , then

2 21
X

X
X

2
2
(1.92)
kxi k
kyi k
hxi , yi i

iI
iI
iI
kxi k2 sin2 i
iI
X
iI

2 21
X

2
2
2
kyi k sin i
hxi , yi i sin i

iI

2 21

X
X
X

2
2
2
2
2
+
kxi k cos i
kyi k cos i
hxi , yi i cos i 0.

iI
iI
iI
(b) If xi , yi H, pi > 0, i {1, . . . , 2n} , then
2n
2 12
2n
2n
X

X
X

pi kxi k2
pi kyi k2
pi hxi , yi i
(1.93)

i=1
n
X
i=1
p2k kx2k k
k=1
i=1
n
X
2
k=1
n
X
n
2 12
X

p2k ky2k k2
p2k hx2k , y2k i

k=1
p2k1 kx2k1 k2
k=1
n
X
p2k1 ky2k1 k2
k=1
2 12

n

X

p2k1 hx2k1 , y2k1 i

( 0) .
k=1
Remark 8. Part (i) of Proposition 15 and the inequality (1.91)

have been obtained by Dragomir and Mond in [2] for the case of scalar
sequences x and y.
Bibliography
[1] S.S. DRAGOMIR and B. MOND, On the superadditivity and monotonicity of
Schwarzs inequality in inner product spaces, Contributions, Macedonian Acad.
of Sci and Arts, 15(2) (1994), 5-22.
[2] S.S. DRAGOMIR and B. MOND, Some inequalities for Fourier coefficients in
inner product spaces, Periodica Math. Hungarica, 32 (3) (1995), 167-172.
[3] S.S. DRAGOMIR and J. SANDOR,

On Bessels and Grams inequalities in
prehilbertian spaces, Periodica Math. Hungarica, 29(3) (1994), 197-205.
[4] F. DEUTSCH, Best Approximation in Inner Product Spaces, CMS Books in
Mathematics, Springer Verlag, New York, Berlin, Heidelberg, 2001.
[5] S. KUREPA, Note on inequalities associated with Hermitian functionals, Glasnik Matematcki, 3(23) (1968), 196-205.
[6] S. KUREPA, On the Buniakowsky-Cauchy-Schwarz inequality, Glasnik
Matematcki, 1(21) (1966), 146-158.
35
CHAPTER 2
Schwarz Related Inequalities

2.1. Introduction
Let H be a linear space over the real or complex number field K.
The functional h, i : H H K is called an inner product on H if it
satisfies the conditions
(i) hx, xi 0 for any x H and hx, xi = 0 iff x = 0;
(ii) hx + y, zi = hx, zi + hy, zi for any , K and x, y, z
H;
(iii) hy, xi = hx, yi for any x, y H.
A first fundamental consequence of the properties (i)-(iii) above, is
the Schwarz inequality:
(2.1)
|hx, yi|2 hx, xi hy, yi ,
for any x, y H. The equality holds in (2.1) if and only if the vectors
x and y are linearly dependent, i.e., there exists a nonzero constant
K so that x = y. p
If we denote kxk := hx, xi, x H, then one may state the following properties
(n) kxk 0 for any x H and kxk = 0 iff x = 0;
(nn) kxk = || kxk for any K and x H;
(nnn) kx + yk kxk+kyk for any x, y H (the triangle inequality);
i.e., kk is a norm on H.
In this chapter we present some classical and recent refinements and
reverse inequalities for the Schwarz and the triangle inequalities. More
precisely, we point out upper bounds or positive lower bounds for the
nonnegative quantities
kxk kyk |hx, yi| , kxk2 kyk2 |hx, yi|2
and
kxk + kyk kx + yk
under various assumptions for the vectors x, y H.
37
38
2. SCHWARZ RELATED INEQUALITIES
If the vectors x, y H are not orthogonal, i.e., hx, yi 6= 0, then

some upper and lower bounds for the supra-unitary quantities
kxk kyk kxk2 kyk2
,
|hx, yi|
|hx, yi|2
under appropriate restrictions for the vectors x and y are provided as
well.
The inequalities obtained by Buzano, Richards, Precupanu and
Moore and their extensions and generalizations for orthonormal families of vectors in both real and complex inner product spaces are presented. Recent results concerning the classical refinement of Schwarz
inequality due to Kurepa for the complexification of real inner product
spaces are also reviewed. Various applications for integral inequalities
including a version of Heisenberg inequality for vector valued functions
in Hilbert spaces are provided as well.
2.2. Inequalities Related to Schwarzs One
2.2.1. Some Refinements. The following result holds [15, Theorem 1] (see also [18, Theorem 2]).
Theorem 10 (Dragomir, 1985). Let (H, h, i) be a real or complex
inner product space. Then

(2.2)
kxk2 kyk2 |hx, yi|2 kyk2 kzk2 |hy, zi|2

2
hx, zi kyk2 hx, yi hy, zi
for any x, y, z H.
Let us consider the mapping
py : H H K,
py (x, z) = hx, zi kyk2 hx, yi hy, zi
for each y H\ {0} .

It is easy to see that py (, ) is a nonnegative Hermitian form and
then on writing Schwarzs inequality
|py (x, z)|2 py (x, x) py (z, z) ,
x, z H
we obtain the desired inequality (2.2).

Remark 9. From (2.2) it follows that [15, Corollary 1] (see also
[18, Corollary 2.1])
1
(2.3) kx + zk2 kyk2 |hx + z, yi|2 2
1
1
kxk2 kyk2 |hx, yi|2 2 + kyk2 kzk2 |hy, zi|2 2
2.2. INEQUALITIES RELATED TO SCHWARZS ONE
39
for every x, y, z H.
Putting z = y in (2.3), we get:
0 kx + yk2 kyk2 |hx + y, yi|2
(2.4)
kxk2 kyk2 |hx, yi|2

and, in particular,
(2.5)
0 kx yk2 kyk2 |hx y, yi|2 kxk2 kyk2 |hx, yi|2
for every x, y H.
Both inequalities (2.4) and (2.5) have been obtained in [15].
We note here that the inequality (2.4) is in fact equivalent to the
following statement

(2.6) sup kx + yk2 kyk2 |hx + y, yi|2 = kxk2 kyk2 |hx, yi|2
K
for each x, y H.
The following corollary may be stated [15, Corollary 2] (see also
[18, Corollary 2.2]):
Corollary 3 (Dragomir, 1985). For any x, y, z H\ {0} we have
the inequality

hx, yi 2 hy, zi 2 hz, xi 2
hx, yi hy, zi hz, xi
+
+
1+2

(2.7)
kxk2 kyk2 kzk2 .
kxk kyk kyk kzk kzk kxk
Proof. By the modulus properties we obviously have

hx, zi kyk2 hx, yi hy, zi |hx, zi| kyk2 |hx, yi| |hy, zi| .
Therefore, by (2.2) we may state that

kxk2 kyk2 |hx, yi|2 kyk2 kzk2 |hy, zi|2
|hx, zi|2 kyk4 2 |hx, yi hy, zi hz, xi| kyk2 + |hx, yi|2 |hy, zi|2 ,
which, upon elementary calculation, is equivalent to (2.7).
Remark 10. If we utilise the elementary inequality a2 + b2 + c2
3abc when a, b, c 0, then one can state the following inequality

hx, yi hy, zi hz, xi hx, yi 2 hy, zi 2 hz, xi 2
+

(2.8) 3
kyk kzk + kzk kxk
kxk2 kyk2 kzk2 kxk kyk
for any x, y, z H\ {0} . Therefore, the inequality (2.7) may be regarded as a reverse inequality of (2.8).
The following refinement of the Schwarz inequality holds [15, Theorem 2] (see also [18, Corollary 1.1]):
40
Theorem 11 (Dragomir, 1985). For any x, y H and e H with

kek = 1, the following refinement of the Schwarz inequality holds:
(2.9)
kxk kyk |hx, yi hx, ei he, yi| + |hx, ei he, yi| |hx, yi| .

Applying the inequality (2.2), we can state that

(2.10) kxk2 |hx, ei|2 kyk2 |hy, ei|2 |hx, yi hx, ei he, yi|2 .
Utilising the elementary inequality for real numbers

(2.11)
m2 n2 p2 q 2 (mp nq)2 ,
we can easily see that
(2.12)
(kxk kyk |hx, ei he, yi|)2

kxk2 |hx, ei|2
kyk2 |hy, ei|2
for any x, y, e H with kek = 1.

Since, by Schwarzs inequality
kxk kyk |hx, ei he, yi|
(2.13)
hence, by (2.10) and (2.12) we deduce the first part of (2.10).

The second part of (2.10) is obvious.
Corollary 4 (Dragomir, 1985). If x, y, e H are such that kek =
1 and x y, then
kxk kyk 2 |hx, ei he, yi| .
(2.14)
Remark 11. Assume that A : H H is a bounded linear operator

on H. For x, e H with kxk = kek = 1, we have by (2.9) that
(2.15) kAyk |hx, Ayi hx, ei he, Ayi| + |hx, ei he, Ayi| |hx, Ayi|
for any y H.
Taking the supremum over x H, kxk = 1 in (2.15) and noting
that kAyk = sup |hx, Ayi| , we deduce the representation
kxk=1
(2.16)
kAyk = sup {|hx, Ayi hx, ei he, Ayi| + |hx, ei he, Ayi|}
kxk=1
for any y H. Finally, on taking the supremum over y H, kyk = 1

in (2.16) we get
(2.17) kAk =
sup
{|hx, Ayi hx, ei he, Ayi| + |hx, ei he, Ayi|}
kyk=1,kxk=1
for any e H, kek = 1, a representation that has been obtained in [15,

Eq. 9].
41
Remark 12. Let (H; h, i) be a Hilbert space. Then for any continuous linear functional f : H K, f 6= 0, there exists, by the Riesz representation theorem a unique vector e H\ {0} such that f (x) = hx, ei
for x H and kf k = kek .
If E is a nonzero linear subspace of H and if we denote by E its
orthogonal complement, i.e., we recall that E := {y H|y x} then
for any x E and y E , by (2.14) we may state that

e
e
,
kxk kyk 2 x,
y,
kxk
kyk
giving, for x, y 6= 0, that
(2.18)
kf k2 2 |hx, ei hy, ei| = 2 |f (x)| |f (y)|
for any x E and y E .

If by kf kE we denote the norm of the functional f restricted to E,
(x)|
i.e., kf kE = supxE\{0} |fkxk
, then, on taking the supremum over x E
and y E in (2.18) we deduce

kf k2 2 kf kE kf kE
(2.19)
for any E a nonzero linear subspace of the Hilbert space H and a given
functional f H \ {0} .
We note that the inequality (2.19) has been obtained in [15, Eq.
10].
2.2.2. A Conditional Inequality. The following result providing a lower bound for the norm product under suitable conditions holds
[19] (see also [18, Theorem 1]):
Theorem 12 (Dragomir-Sandor, 1986). Let x, y, a, b H, where
(H; h, i) is an inner product space, be such that
(2.20)
kak2 2 Re hx, ai
and
kyk2 2 Re hy, bi
holds true. Then

(2.21) kxk kyk 2 Re hx, ai kak2
21
2 Re hy, bi kbk2
12
+ |hx, yi hx, bi ha, yi + ha, bi| .

42
Observe that
(2.22) |hx, yi hx, bi ha, yi + ha, bi|
= |hx a, y bi|2
kx ak2 ky bk2

= kxk2 2 Re hx, ai kak2
kyk2 2 Re hy, bi kbk2 .
Applying the elementary inequality (2.11) we have

h
1 i2
2 2
2
(2.23)
kxk 2 Re hx, ai kak

h
1 i2
2
2 2
kyk 2 Re hy, bi kbk
h
1 i
1
kxk kyk 2 Re hx, ai kak2 2 2 Re hy, bi kbk2 2 .
Since
1
0 2 Re hx, ai kak2 2 kxk
1
0 2 Re hy, bi kbk2 2 kyk
and
hence
kxk kyk 2 Re hx, ai kak2
12
2 Re hy, bi kbk2
12
and by (2.22) and (2.23) we deduce the desired result (2.21).

Remark 13. As pointed out in [19], if we consider a = hx, ei e,
b = hy, ei e with e H, kek = 1, then the condition (2.20) is obviously
satisfied and the inequality (2.21) becomes
(2.24)
kxk kyk |hx, ei he, yi| + |hx, yi hx, ei he, yi|

( |hx, yi|),
which is the refinement of the Schwarz inequality incorporated in (2.9).
For vectors located in a closed ball centered at 0 and of radius 2,

one can state the following corollary as well [18, Corollary 1.2].
Corollary 5. Let x, y H such that kxk , kyk 2. Then

(2.25) kxk kyk |hx, yi|2 2 kxk2
12
2 kyk2
12

+ |hx, yi| 1 kxk2 kyk2 + |hx, yi|2 .
Proof. Follows by Theorem 12 on choosing a = hx, yi y, b =
hy, xi x. We omit the details.
43
2.2.3. A Refinement for Orthonormal Families. The following result provides a generalisation for a refinement of the Schwarz
inequality incorporated in (2.9) [15, Theorem 3] (see also [8, Theorem]
or [18, Theorem 3]):
Theorem 13 (Dragomir, 1985). Let (H; h, i) be an inner product
space over the real or complex number field K and {ei }iI an orthonormal family in I. For any F a nonempty finite part of I we have the
following refinement of Schwarzs inequality:

X
X

(2.26)
kxk kyk hx, yi
hx, ei i hei , yi +
|hx, ei i hei , yi|

iF
iF

X

X

hx, yi
hx, ei i hei , yi +
hx, ei i hei , yi

iF
iF
|hx, yi| ,
where x, y H.
We apply the Schwarz inequality to obtain
*
+2

X
X

(2.27) x
hx, ei i ei , y
hy, ei i ei

iF
iF

2
2

X
X

hx, ei i ei y
hy, ei i ei .
x

iF
iF
Since a simple calculation with orthonormal vectors shows that

2

X
X

hx, ei i ei = kxk2
|hx, ei i|2 ,
x

iF
iF

2

X
X

hy, ei i ei = kyk2
|hy, ei i|2 ,
y

iF
iF
and
*
x
+
X
iF
hx, ei i ei , y
X
iF
hy, ei i ei
= hx, yi
X
iF
hx, ei i hei , yi ,
44
hence (2.27) is equivalent to

2

X

(2.28) hx, yi
hx, ei i hei , yi

iF
!
kxk2
|hx, ei i|2
kyk2
iF
|hy, ei i|2
iF
for any x, y H.
Further, we need the following Aczel type inequality
(2.29)
2i
!2
!
X
2i
iF
iF
i i
iF
P
P
provided that 2 iF 2i and 2 iF 2i , where , , i , i R,
i F.
For an Aczel inequality that holds under slightly weaker conditions
and a different proof based on polynomials, see [26, p. 57].
For the sake of completeness, we give here a direct proof of (2.29).
Utilising the elementary inequality (2.11), we can write
! 12 2
(2.30) 2
2i
! 12 2
iF
2i
iF
||
2i
iF
2i
21
and ||
iF
2i
12
! 12
||
X
iF
2i
, then
! 12
X
iF
2i
X
iF
iF
Since ||
! 12 2
! 12
2i
45
Therefore, by the Cauchy-Bunyakovsky-Schwarz inequality, we have

that

! 12
! 12
! 12
! 12

X
X
X
X

||
2i
2i = ||
2i
2i

iF
iF
iF
iF

X

||
i i

iF

X

= ||
i i

iF

X

i i ,

iF
showing that
(2.31)
! 12 2
! 12
||
2i
iF
2i
!2
i i
iF
iF
and then, by (2.30) and (2.31) we deduce the desired result (2.29).
By Bessels inequality we obviously have that
X
X
kxk2
|hx, ei i|2
and
kyk2
|hy, ei i|2 ,
iF
iF
therefore, on applying the inequality (2.29) we deduce that

!
(2.32)
kxk2
|hx, ei i|2
!
kyk2
iF
|hy, ei i|2
iF
!2
kxk kyk
|hx, ei i hei , yi|
iF
Since kxk kyk iF |hx, ei i hei , yi| 0, hence by (2.28) and (2.32) we
deduce the first part of (2.26).
The second and third parts are obvious.
When the vectors are orthogonal, the following result may be stated
[8] (see also [18, Corollary 3.1]).
46
Corollary 6. If {ei }iI is an orthonormal family in (H, h, i) and

x, y H with x y, then we have the inequality:

X
X

(2.33)
kxk kyk
hx, ei i hei , yi +
|hx, ei i hei , yi|

iF
iF

X

2
hx, ei i hei , yi ,

iF
for any nonempty finite part of I.

2.3. Kurepa Type Refinements for the Schwarz Inequality
2.3.1. Kurepas Inequality. In 1960, N.G. de Bruijn proved the
following refinement of the celebrated Cauchy-Bunyakovsky-Schwarz
(CBS) inequality for a sequence of real numbers and the second of
complex numbers, see [2] or [9, p. 48]:
Theorem 14 (de Bruijn, 1960). Let (a1 , . . . , an ) be an ntuple of
real numbers and (z1 , . . . , zn ) an ntuple of complex numbers. Then
n
#
n
2
" n
n

X

X
X
X
1

2
2
2
(2.34)
a
|zk | +
zk
ak zk

2 k=1 k k=1
k=1
k=1
!
n
n
X
X
2
ak
|zk |2 .
k=1
k=1
Equality holds in (2.34) if and only if, for k P

{1, . . . , n} , ak =
n
2
2
Re (zk ) , where is a complex number such that
k=1 zn is a nonnegative real number.
In 1966, in an effort to extend this result to inner products, Kurepa
[25] obtained the following refinement for the complexification of a real
inner product space (H; h, i) :
Theorem 15 (Kurepa, 1966). Let (H; h, i) be a real inner product
space and (HC , h, iC ) its complexification. For any a H and z HC
we have the inequality:

1
(2.35)
|hz, aiC |2 kak2 kzk2C + |hz, ziC |
2

kak2 kzk2C .
To be comprehensive, we define in the following the concept of
complexification for a real inner product space.
Let H be a real inner product space with the scalar product h, i
and the norm kk . The complexification HC of H is defined as a complex
2.3. KUREPA TYPE REFINEMENTS FOR THE SCHWARZ INEQUALITY 47
linear space H H of all ordered pairs (x, y) (x, y H) endowed with

the operations
(x, y) + (x0 , y 0 ) := (x + x0 , y + y 0 ) ,
x, x0 , y, y 0 H;
( + i ) (x, y) := (x y, x + y) ,
x, y H and , R.
On HC one can canonically consider the scalar product h, iC defined

by:
hz, z 0 iC := hx, x0 i + hy, y 0 i + i [hy, x0 i hx, y 0 i]
where z = (x, y) , z 0 = (x0 , y 0 ) HC . Obviously,
kzk2C = kxk2 + kyk2 ,
where z = (x, y) .
The conjugate of a vector z = (x, y) HC is defined by z :=
(x, y) .
It is easy to see that the elements of HC under defined operations
behave as formal complex combinations x+iy with x, y H. Because
of this, we may write z = x + iy instead of z = (x, y) . Thus, z = x iy.
2.3.2. A Generalisation of Kurepas Inequality. The following lemma is of interest [6].
Lemma 2. Let f : [0, 2] R given by
(2.36)
f () = sin2 + 2 sin cos + cos2 ,
where , , R. Then
(2.37)
sup f () =
[0,2]
1
1
1
( + ) + ( )2 + 4 2 2 .
2
2
Proof. Since
1 cos 2
1 + cos 2
sin2 =
, cos2 =
, 2 sin cos = sin 2,
2
2
hence f may be written as
1
1
(2.38)
f () = ( + ) + ( ) cos 2 + sin 2.
2
2
If = 0, then (2.38) becomes
f () =
1
1
( + ) + ( ) cos 2.
2
2
Obviously, in this case

sup f () =
[0,2]
1
1
( + ) + | | = max {, } .
2
2
48
If 6= 0, then (2.38) becomes

1
( )
f () = ( + ) + sin 2 +
cos 2 .
2

Let 2 , 2 for which tan =
. Then f can be written as
2
f () =
( + ) +
sin (2 + ) .
2
cos
For this function, obviously

(2.39)
sup f () =
[0,2]
1
||
( + ) +
.
2
|cos |
Since
sin2
( )2
=
,
cos2
4 2
hence,

1
( )2 + 4 2 2
1
=
,
|cos |
2 ||
and from (2.39) we deduce the desired result (2.37).
The following result holds [6].
Theorem 16 (Dragomir, 2004). Let (H; h, i) be a complex inner
product space. If x, y, z H are such that
(2.40)
Im hx, zi = Im hy, zi = 0,
then we have the inequality:

(2.41)
Re2 hx, zi + Re2 hy, zi

= |hx + iy, zi|2

h
i 12
1
2
2
2
2 2
2
kxk + kyk + kxk kyk 4 Re hx, yi

kzk2
2

kxk2 + kyk2 kzk2 .
Proof. Obviously, by (2.40), we have

hx + iy, zi = Re hx, zi + i Re hy, zi
and the first part of (2.41) holds true.
Now, let [0, 2] be such that hx + iy, zi = ei |hx + iy, zi| .
Then

|hx + iy, zi| = ei hx + iy, zi = ei (x + iy) , z .
2.3. KUREPA TYPE REFINEMENTS FOR THE SCHWARZ INEQUALITY 49
Utilising the above identity, we can write:

|hx + iy, zi| = Re ei (x + iy) , z
= Re h(cos i sin ) (x + iy) , zi
= Re hcos x + sin y i sin x + i cos y, zi
= Re hcos x + sin y, zi + Im hsin x cos y, zi
= Re hcos x + sin y, zi + sin Im hx, zi cos Im hy, zi
= Re hcos x + sin y, zi ,
and for the last equality we have used the assumption (2.40).
Taking the square and using the Schwarz inequality for the inner
product h, i , we have
|hx + iy, zi|2 = [Re hcos x + sin y, zi]2
(2.42)
kcos x + sin yk2 kzk2 .

On making use of Lemma 2, we have
sup kcos x + sin yk2
[0,2]
= sup

kxk2 cos2 + 2 Re hx, yi sin cos + kyk2 sin 2
[0,2]
1
=
2

h
i 21
2
2 2
2
2
2
kxk + kyk + kxk kyk + 4 Re hx, yi
and the first inequality in (2.41) is proved.

Observe that
2
kxk2 kyk2 + 4 Re2 hx, yi
2

= kxk2 + kyk2 4 kxk2 kyk2 Re2 hx, yi
2
kxk2 + kyk2
and the last part of (2.41) is proved.
Remark 14. Observe that if (H, h, i) is a real inner product space,
then for any x, y, z H one has:
(2.43)
hx, zi2 + hy, zi2

h
io 12
1n
2
2
2
2 2
2
kxk + kyk + kxk kyk + 4 hx, yi

kzk2
2

kxk2 + kyk2 kzk2 .
50
Remark 15. If H is a real space, h, i the real inner product, HC

its complexification and h, iC the corresponding complexification for
h, i, then for x, y H and w := x + iy HC and for e H we have
Im hx, eiC = Im hy, eiC = 0,
kwk2C = kxk2 + kyk2 ,
|hw, wi
C | = kxk2 kyk2
2
+ 4 hx, yi2 ,
where w = x iy HC .
Applying Theorem 16 for the complex space HC and complex inner
product h, iC , we deduce

1
(2.44)
|hw, eiC |2 kek2 kwk2C + |hw, wi
C | kek2 kwk2C ,
2
which is Kurepas inequality (2.35).
Corollary 7. Let x, y, z be as in Theorem 16. In addition, if
Re hx, yi = 0, then
2
1
(2.45)
Re hx, zi + Re2 hy, zi 2 kzk max {kxk , kyk} .
Remark 16. If H is a real space and h, i a real inner product on
H, then for any x, y, z H with hx, yi = 0 we have

1
(2.46)
hx, zi2 + hy, zi2 2 kzk max {kxk , kyk} .
2.3.3. A Related Result. Utilising Lemma 2, we may state and
prove the following result as well.
Theorem 17 (Dragomir, 2004). Let (H, h, i) be a real or complex
inner product space. Then we have the inequalities:

h
2
1
2
2
(2.47)
|hv, ti| + |hw, ti| + |hv, ti|2 |hw, ti|2
2
1 o
+ 4 (Re hv, ti Re hw, ti + Im hv, ti Im hw, ti)2 2

h
i 12
1
2
2
2
2
2 2
2
ktk kvk + kwk + kvk kwk + 4 Re (v, w)
2

kvk2 + kwk2 ktk2 ,
for all v, w, t H.
Proof. Observe that, by Schwarzs inequality
(2.48)
|(cos v + sin w, z)|2 kcos v + sin wk2 kzk2
for any [0, 2] .
2.4. REFINEMENTS OF BUZANOS AND KUREPAS INEQUALITIES
51
Since
I () := kcos v + sin wk2
= cos2 kvk2 + 2 Re (v, w) sin cos + kwk2 sin2 ,
hence, as in Theorem 16,

h
i 12
1
2
2
2
2 2
2
.
sup I () =
kvk + kwk + kvk kwk + 4 Re (v, w)
2
[0,2]
Also, denoting
J () := |cos hv, zi + sin hw, zi|
i
= cos |hv, zi| + 2 sin cos Re hv, zi hw, zi + sin2 |hw, zi|2 ,
2
then, on applying Lemma 2, we deduce that

1
sup J () =
|hv, ti|2 + |hw, ti|2
2
[0,2]
h
h
ii 12
2
2 2
2
+ |hv, ti| |hw, ti| + 4 Re hv, zi hw, zi
and, since
h
i
Re hv, zi hw, zi = Re hv, ti Re hw, ti + Im hv, ti Im hw, ti ,
hence, on taking the supremum in the inequality (2.48), we deduce the
desired inequality (2.47).
Remark 17. In the real case, (2.47) provides the same inequality
we obtained in (2.43).
In the complex case, if we assume that v, w, t H are such that
Re hv, ti Re hw, ti = Im hv, ti Im hw, ti ,
then (2.47) becomes:

(2.49) max |hv, ti|2 , |hw, ti|2

h
i 21
1
2
2
2
2
2 2
2
ktk kvk + kwk + kvk kwk + 4 Re (v, w)
.
2
2.4. Refinements of Buzanos and Kurepas Inequalities
2.4.1. Introduction. In [3], M.L. Buzano obtained the following
extension of the celebrated Schwarzs inequality in a real or complex
1
(2.50)
|ha, xi hx, bi| [kak kbk + |ha, bi|] kxk2 ,
2
52
for any a, b, x H.
It is clear that for a = b, the above inequality becomes the standard
Schwarz inequality
|ha, xi|2 kak2 kxk2 ,
(2.51)
a, x H;
with equality if and only if there exists a scalar K (K = R or C)

such that x = a.
As noted by M. Fujii and F. Kubo in [21], where they provided
a simple proof of (2.50) by utilising orthogonal projection arguments,
the case of equality holds in (2.50) if

ha,bi
a
b
, when ha, bi =
6 0
kak
|ha,bi| kbk
x=

a + b ,
when ha, bi = 0,
kak
kbk
where , K.
It might be useful to observe that, out of (2.50), one may get the
following discrete inequality:

n
n

X
X

(2.52)
p i ai x i
pi xi bi

i=1
i=1

! 12 n
n
n
n
X

X
X
X
1

2
2
pi |ai |
pi |bi |
+
p i ai b i
pi |xi |2 ,

2
i=1
i=1
i=1
i=1
where pi 0, ai , xi , bi C, i {1, . . . , n} .
If one takes in (2.52) bi = ai for i {1, . . . , n} , then one obtains
(2.53)
n

#
" n
n
n
n
X
1 X
X
X
X

2
2
p i ai x i
p i ai x i
pi |ai | +
p i ai
pi |xi |2 ,

2

i=1
i=1
i=1
i=1
i=1
for any pi 0, ai , xi , bi C, i {1, . . . , n} .

Note that, if xi , i {1, . . . , n} are real numbers, then out of (2.53),
we may deduce the de Bruijn refinement of the celebrated CauchyBunyakovsky-Schwarz inequality [2]

2

#
" n
n
n
n
X

X

1X 2 X

2
(2.54)
pi xi zi
pi xi
pi |zi | +
pi zi2 ,

2
i=1
i=1
i=1
i=1
where zi C, i {1, . . . , n} . In this way, Buzanos result may be

regarded as a generalisation of de Bruijns inequality.
Similar comments obviously apply for integrals, but, for the sake of
brevity we do not mention them here.
53
The aim of the present section is to establish some related results

as well as a refinement of Buzanos inequality for real or complex inner
product spaces. An improvement of Kurepas inequality for the complexification of a real inner product and the corresponding applications
for discrete and integral inequalities are also provided.
2.4.2. Some Buzano Type Inequalities. The following result
may be stated [16].
space over the real or complex number field K. For all K\ {0} and
x, a, b H, 6= 0, one has the inequality

ha, xi hx, bi ha, bi

(2.55)

kxk2

kbk
| 1|2 |ha, xi|2 + kxk2 kak2 |ha, xi|2 .

|| kxk
The case of equality holds in (2.55) if and only if there exists a scalar
K so that
(2.56)
ha, xi
x = a + b.
kxk2

Using Schwarzs inequality, we have that

2
2

ha, xi

ha, xi

kbk2
(2.57)
a,
b
a

kxk2
kxk2
and since

2
2
ha, xi

|ha, xi|2
2 |ha, xi|
2

2
2 x a = ||
2
2 Re + kak

kxk
kxk
kxk
| 1|2 |ha, xi|2 + kxk2 kak2 |ha, xi|2
=
kxk2
and

ha, xi
x a, b =
kxk2
kxk2
hence by (2.55) we deduce the desired inequality (2.55).
The case of equality is obvious from the above considerations related
to the Schwarzs inequality (2.51).
54
Remark 18. Using the continuity property of the modulus, i.e.,

||z| |u|| |z u| , z, u K, we have:

|ha, xi hx, bi| |ha, bi| ha, xi hx, bi ha, bi

.
(2.58)
2
2

||
kxk
kxk
Therefore, by (2.55) and (2.58), one may deduce the following double
inequality:

1
kbk
(2.59)
|ha, bi|
||
kxk
h
1 ii
| 1|2 |hx, ai|2 + kxk2 kak2 |ha, xi|2 2
|ha, xi hx, bi|
kxk2

1
kbk
|ha, bi| +
||
kxk
h
1i
2
2
2
2
2 2
| 1| |hx, ai| + kxk kak |hx, ai|
,
for each K\ {0} , a, b, x H and x 6= 0.

It is obvious that, out of (2.55), we can obtain various particular
inequalities. We mention in the following a class of these which is
related to Buzanos result (2.50) [16].
Corollary 8 (Dragomir, 2004). Let a, b, x H, x 6= 0 and K
with || = 1, Re 6= 1. Then we have the inequality:

kbk

kak
(2.60)
kxk2

1+
2 1 + Re
and, in particular, for = 1, the inequality:

ha, xi hx, bi ha, bi kak kbk

(2.61)
.
kxk2
2
2
Proof. It follows by Theorem 18 on choosing = 1 + and we
omit the details.
Remark 19. Using the continuity property of modulus, we get from
(2.60) that:
|ha, bi| + kak kbk
|ha, xi hx, bi|

,
|| = 1, Re 6= 1,
2
2 1 + Re
kxk
which provides, as the best possible inequality, the above result due to
Buzano (2.50).
55
Remark 20. If the space is real, then the inequality (2.55) is obviously equivalent to:
(2.62)
1
ha, bi
kbk
( 1)2 ha, xi2 + kxk2 kak2 ha, xi2 2
|| kxk
ha, xi hx, bi
kxk2
1
ha, bi
kbk
+
( 1)2 ha, xi2 + kxk2 kak2 ha, xi2 2
|| kxk
for any R\ {0} and a, b, x H, x 6= 0.

If in (2.62) we take = 2, then we get
1
[ha, bi kak kbk] kxk2 ha, xi hx, bi
2
1
[ha, bi + kak kyk] kxk2 ,
2
which apparently, as mentioned by T. Precupanu in [29], has been obtained independently of Buzano, by U. Richard in [30].
In [28], Pecaric gave a simple direct proof of (2.63) without mentioning the work of either Buzano or Richard, but tracked down the
result, in a particular form, to an earlier paper due to C. Blatter [1].
(2.63)
Obviously, the following refinement of Buzanos result may be stated

[16].
Corollary 9 (Dragomir, 2004). Let (H; h, i) be a real or complex
inner product space and a, b, x H. Then

1
1
2
(2.64) |ha, xi hx, bi| ha, xi hx, bi ha, bi kxk + |ha, bi| kxk2
2
2
1
[kak kbk + |ha, bi|] kxk2 .
2
Proof. The first inequality in (2.64) follows by the triangle inequality for the modulus || . The second inequality is merely (2.61) in
which we added the same quantity to both sides.
Remark 21. For = 1, we deduce from (2.55) the following
inequality:

ha, xi hx, bi

1

kbk kxk2 kak2 |ha, xi|2 2
(2.65)
ha,
bi
kxk2
kxk
for any a, b, x H with x 6= 0.
56
If the space is real, then (2.65) is equivalent to

(2.66)
1
kbk
kxk2 kak2 |ha, xi|2 2
kxk
ha, xi hx, bi
kxk2
1
kbk
kxk2 kak2 |ha, xi|2 2 + ha, bi ,

kxk
ha, bi
which is similar to Richards inequality (2.63).

2.4.3. Applications to Kurepas Inequality. In 1960, N.G. de
Bruijn [2] obtained the following refinement of the Cauchy-BunyakovskySchwarz inequality:

#
2
" n
n
n
n
X

1 X 2 X 2 X 2

(2.67)
ai zi
ai
|zi | +
zi ,

2
i=1
i=1
i=1
i=1
provided that ai are real numbers while zi are complex for each i
{1, ..., n} .
In [25], S. Kurepa proved the following generalisation of the de
Bruijn result:
Theorem 19 (Kurepa, 1966). Let (H; h, i) be a real inner product
space and (HC , h, iC ) its complexification. Then for any a H and
z HC , one has the following refinement of Schwarzs inequality

1
kak2 kzk2C + |hz, ziC | kak2 kzk2C ,
2
where z denotes the conjugate of z HC .
(2.68)
|ha, ziC |2
As consequences of this general result, Kurepa noted the following

integral, respectively, discrete inequality:
Corollary 10 (Kurepa, 1966). Let (S, , ) be a positive measure
space and a, z L2 (S, , ) , the Hilbert space of complex-valued 2
integrable functions defined on S. If a is a real function and z is a
complex function, then
Z
2

(2.69) a (t) z (t) d (t)
S
Z

Z
Z

1
2
2
2

a (t) d (t)
|z (t)| d (t) + z (t) d (t) .
2 S
S
S
57
Corollary 11 (Kurepa, 1966). If a1 , . . . , an are real numbers,

z1 , . . . , zn are complex numbers and (Aij ) is a positive definite real matrix of dimension n n, then
n
2
n
#
" n
n
X

X
X
X
1

(2.70)
Aij ai zj
Aij ai aj
Aij zi zj +
Aij zi zj .

2 i,j=1
i,j=1
i,j=1
i,j=1
The following refinement of Kurepas result may be stated [16].
Theorem 20 (Dragomir, 2004). Let (H; h, i) be a real inner product space and (HC , h, iC ) its complexification. Then for any e H
and w HC , one has the inequality:

1

1
2
2
2
(2.71)
|hw, eiC | hw, eiC hw, wi
C kek + |hw, wi
C | kek2
2
2

1
kek2 kwk2C + |hw, wi
C| .
2
If we apply Corollary 11 for (HC , h, iC ) and x = e H, a = w and
b = w,
then we have
(2.72)
|hw, eiC he, wi

C|

1
1
2
C kek + |hw, wi
C | kek2
hw, eiC he, wi
C hw, wi
2
2
1
C + |hw, wi
C |] .
kek2 [kwkC kwk
2
Now, if we assume that w = (x, y) HC , then, by the definition of

h, iC , we have
hw, eiC = h(x, y) , (e, 0)iC
= hx, ei + hy, 0i + i [hy, ei hx, 0i]
= he, xi + i he, yi ,
he, wi
C = h(e, 0) , (x, y)iC
= he, xi + h0, yi + i [h0, xi he, yi]
= he, xi + i he, yi = hw, eiC
and
kwk
2C = kxk2 + kyk2 = kwk2C .
Therefore, by (2.72), we deduce the desired result (2.71).
58
Denote by `2 (C) the Hilbert space of all complex

P sequences z =
(zi )iN with the property that for i 0 with i=1 i = 1 we have
P
2
i=1 i |zi | < . If a = (ai )iN is a sequence of real numbers such
that a `2 (C) , then for any z `2 (C) we have the inequality:

2
X

(2.73)
i ai zi

i=1

!2

X

X
X
X
X

1
1

i a2i
i zi2 +
i a2i
i zi2

i ai zi

2 i=1
i=1
2 i=1
i=1
i=1

#
"
X

1X 2 X

i ai
i |zi |2 +
i zi2 .

2
i=1
i=1
i=1
L2
Similarly, if by
(S, , ) we understand the Hilbert space of all
complex-valued functions f : S RC with the property that for the
measurable function 0 with S (t) d (t) = 1 we have
Z
(t) |f (t)|2 d (t) < ,
S
then for a real function a L2 (S, , ) and any f L2 (S, , ) , we

have the inequalities
Z
2

(2.74) (t) a (t) f (t) d (t)
S
Z
2

(t) a (t) f (t) d (t)
S

Z
Z

1
2
2
(t) f (t) d (t) (t) a (t) d (t)
2 S
S
Z
Z

1
+ (t) f 2 (t) d (t) (t) a2 (t) d (t)
2 S
S
Z
1
(t) a2 (t) d (t)

2 S
Z

Z

2
2

(t) |f (t)| d (t) + (t) f (t) d (t) .

S
2.5. Inequalities for Orthornormal Families

2.5.1. Introduction. In [3], M.L. Buzano obtained the following
extension of the celebrated Schwarzs inequality in a real or complex
2.5. INEQUALITIES FOR ORTHORNORMAL FAMILIES
59

|ha, xi hx, bi|
(2.75)
1
[kak kbk + |ha, bi|] kxk2 ,
2
for any a, b, x H.
It is clear that the above inequality becomes, for a = b, the Schwarzs
inequality
|ha, xi|2 kak2 kxk2 ,
(2.76)
a, x H;
in which the equality holds if and only if there exists a scalar K

(R, C) so that x = a.
As noted by T. Precupanu in [29], independently of Buzano, U.
Richard [30] obtained the following similar inequality holding in real
inner product spaces:
1
kxk2 [ha, bi kak kbk] ha, xi hx, bi
2
1
kxk2 [ha, bi + kak kbk] .
2
(2.77)
The main aim of the present section is to obtain similar results for
families of orthonormal vectors in (H; h, i) , real or complex space,
that are naturally connected with the celebrated Bessel inequality and
improve the results of Busano, Richard and Kurepa.
2.5.2. A Generalisation for Orthonormal Families. We say
that the finite family {ei }iI (I is finite) of vectors is orthonormal if
hei , ej i = 0 if i, j I with i 6= j and kei k = 1 for each i I. The
following result may be stated [11]:
Theorem 21 (Dragomir, 2004). Let (H; h, i) be an inner product space over the real or complex number field K and {ei }iI a finite
orthonormal family in H. Then for any a, b H, one has the inequality:

X
1
1

(2.78)
ha, ei i hei , bi ha, bi kak kbk .

2
2
iI
The case of equality holds in (2.78) if and only if
(2.79)
X
iI
1
ha, ei i ei = a +
2
X
iI
!
1
b
.
ha, ei i hei , bi ha, bi
2
kbk2
60
It is well known that, for e 6= 0 and f H, the following identity

holds:
2

kf k2 kek2 |hf, ei|2
hf, ei e

.
(2.80)
= f
kek2
kek2
Therefore, in Schwarzs inequality
|hf, ei|2 kf k2 kek2 ,
(2.81)
f, e H;
the case of equality, for e 6= 0, holds if and only if

f=
hf, ei e
.
kek2
P
Let f := 2 iI ha, ei i ei a and e := b. Then, by Schwarzs inequality
(2.81), we may state that
*
2
+2
X

X

(2.82)
ha, ei i ei a, b 2
ha, ei i ei a kbk2
2

iI
iI
with equality, for b 6= 0, if and only if

*
+
X
X
b
(2.83)
2
ha, ei i ei a = 2
ha, ei i ei a, b
2.
kbk
iI
iI
Since
+
*
2
ha, ei i ei a, b
iI
=2
iI
and

2
X

ha, ei i ei a
2

iI

2
*
+
X

X

= 4
ha, ei i ei 4 Re
ha, ei i ei , a + kak2

iI
iI
X
X
=4
|ha, ei i|2 4
|ha, ei i|2 + kak2
iI
2
iI
= kak ,
hence by (2.82) we deduce the desired inequality (2.78).
Finally, as (2.79) is equivalent to
!
X
X
a
1
b
ha, ei i ei =
2,
2
2
kbk
iI
iI
61
hence the equality holds in (2.78) if and only if (2.79) is valid.

The following result is well known in the literature as Bessels
inequality
X
(2.84)
|hx, ei i|2 kxk2 , x H,
iI
where, as above, {ei }iI is a finite orthonormal family in the inner

product space (H; h, i) .
If one chooses a = b = x in (2.78), then one gets the inequality

X
1
1

2
2
|hx, ei i| kxk kxk2 ,

2
2
iI
which is obviously equivalent to Bessels inequality (2.84). Therefore,

the inequality (2.78) may be regarded as a generalisation of Bessels
inequality as well.
Utilising the Bessel and Cauchy-Bunyakovsky-Schwarz inequalities,
one may state that

"
# 12
X

X
X

2
2
(2.85)
ha, ei i hei , bi
|ha, ei i|
|hb, ei i|
kak kbk

iI
iI
iI
A different refinement of the inequality between the first and the

last term in (2.85) is incorporated in the following [11]:
Corollary 12 (Dragomir, 2004). With the assumption of Theorem 21, we have

X
X
1
1

(2.86)
ha, ei i hei , bi
ha, ei i hei , bi ha, bi + |ha, bi|

2
2
iI
iI
1
[kak kbk + |ha, bi|]
2
kak kbk .
Remark 22. If the space (H; h, i) is real, then, obviously, (2.78)
is equivalent to:
X
1
1
(2.87)
(ha, bi kak kbk)
ha, ei i hei , bi [kak kbk + ha, bi] .
2
2
iI
Remark 23. It is obvious that if the family comprises of only a
x
, x H, x 6= 0, then from (2.86) we recapture
single element e = kxk
the refinement of Buzanos inequality incorporated in (2.75) while from
(2.87) we deduce Richards result from (2.77).
62
The following corollary of Theorem 21 is of interest as well [11]:

Corollary 13 (Dragomir, 2004). Let {ei }iI be a finite orthonormal family in (H; h, i) . If x, y H\ {0} are such that there exists the
constants mi , ni , Mi , Ni R, i I such that:
(2.88)
1 mi
Re hx, ei i Re hy, ei i
Mi 1,
kxk
kyk
iI
1 ni
Im hx, ei i Im hy, ei i
Ni 1,
kxk
kyk
iI
and
(2.89)
then
(2.90)
(mi + ni ) 1
iI
X
Re hx, yi
1+2
(Mi + Ni ) .
kxk kyk
iI

Using Theorem 21 and the fact that for any complex number z,
|z| |Re z| , we have

X

1

(2.91)
Re [hx, ei i hei , yi] Re hx, yi

2
iI

X

1

hx, ei i hei , yi hx, yi

2
iI
1
kxk kyk .
2
Since
Re [hx, ei i hei , yi] = Re hx, ei i Re hy, ei i + Im hx, ei i Im hy, ei i ,
hence by (2.91) we have:
(2.92)
1
1
kxk kyk + Re hx, yi
2X
2
X
Re hx, ei i Re hy, ei i +
Im hx, ei i Im hy, ei i
iI
iI
1
1
kxk kyk + Re hx, yi .
2
2
Utilising the assumptions (2.88) and (2.89), we have
X
X Re hx, ei i Re hy, ei i X
(2.93)
mi
Mi
kxk
kyk
iI
iI
iI
63
and
(2.94)
X
iI
ni
X Im hx, ei i Im hy, ei i
iI
kxk kyk
Ni .
iI
Finally, on making use of (2.92) (2.94), we deduce the desired result

(2.90).
Remark 24. By Schwarzs inequality, is it obvious that, in general,
1
Re hx, yi
1.
kxk kyk
P
Consequently, the left inequality in (2.90) is of interest when
iI (mi + ni ) >
P
0, while the right inequality in (2.90) is of interest when iI (Mi + Ni ) <
0.
2.5.3. Refinements of Kurepas Inequality. The following result holds [11].
Theorem 22 (Dragomir, 2004). Let {ej }jI be a finite orthonormal
family in the real inner product space (H; h, i) . Then for any w
HC , where (HC ; h, iC ) is the complexification of (H; h, i) , one has the
following Bessels type inequality:

X
X
1
1

(2.95)
C + |hw, wi
C|
hw, ej i2C
hw, ej i2C hw, wi

2
2
jI
jI

1
kwk2C + |hw, wi
C | kwk2C .
2
Define fj HC , fj := (ej , 0) , j I. For any k, j I we have
hfi , fj iC = h(ek , 0) , (ej , 0)iC = hek , ej i = kj ,

therefore {fj }jI is an orthonormal family in (HC ; h, iC ) .
If we apply Theorem 21 for (HC ; h, iC ) , a = w, b = w,
we may
write:

X
1
1

(2.96)
hw, ej iC hej , wi
C hw, wi
C kwkC kwk
C.

2
2
jI
However, for w := (x, y) HC , we have w = (x, y) and
hej , wi
C = h(ej , 0) , (x, y)iC = hej , xi i hej , yi = hej , xi + i hej , yi
and
hw, ej iC = h(x, y) , (ej , 0)iC = hej , xi i hej , yi = hx, ej i + i hej , yi
64
for any j I. Thus hej , wi

= hw, ej i for each j I and since
kwkC = kwk
C = kxk2 + kyk2
21
we get from (2.96) that

X
1
1

(2.97)
C kwk2C .
hw, ej i2C hw, wi

2

2
jI
Now, observe that the first inequality in (2.95) follows by the triangle
inequality, the second is an obvious consequence of (2.97) and the last
one is derived from Schwarzs result.
Remark 25. If the family {ej }jI contains only a single element
x
e = kxk
, x H, x 6= 0, then from (2.95) we deduce (2.72), which, in
its turn, provides a refinement of Kurepas inequality (2.68).
2.5.4. An Application for L2 [, ]. It is well known that in
the Hilbert space L2 [, ] of all functions f : [, ]R C with the
property that f is Lebesgue measurable on [, ] and |f (t)|2 dt <

, the set of functions

1
1
1
1
1
, cos t, sin t, . . . , cos nt, sin nt, . . .
2
is orthonormal.
If by trig t, we denote either sin t or cos t, t [, ] , then on using
the results from Sections 2.5.2 and 2.5.3, we may state the following
inequality:
(2.98)

Z
n Z
1 X

f (t) trig (kt) dt
g (t) trig (kt) dt

k=1
2
Z

1
f (t) g (t)dt
2
Z
Z
1
2
|f (t)| dt
|g (t)|2 dt,
4
where all trig (kt) is either sin kt or cos kt, k {1, . . . , n} and f
L2 [, ] .
This follows by Theorem 21.
2.6. GENERALIZATIONS OF PRECUPANU S INEQUALITY
65
If one uses Corollary 12, then one can state the following chain of
inequalities
(2.99)

Z
n Z
1 X

f (t) trig (kt) dt
g (t) trig (kt) dt

k=1 n Z
Z
1 X

f (t) trig (kt) dt
g (t) trig (kt) dt

k=1

Z
Z
1

1

f (t) g (t)dt +
f (t) g (t)dt
2
2
" Z
#
12 Z
Z
|f (t)|2 dt
|g (t)|2 dt
+
f (t) g (t)dt
2
1
Z

Z
2
2
|f (t)| dt
|g (t)|2 dt ,
where f L2 [, ] .
Finally, by employing Theorem 22, we may state:

2
n Z
X

f (t) trig (kt) dt

k=1

2
Z
n Z
1 X
1 Z

1 2

2

f (t) trig (kt) dt
f (t) dt +
f (t) dt

2
2
k=1
Z
Z
Z

1
|f (t)|2 dt +
f 2 (t) dt
|f (t)|2 dt,
where f L2 [, ] .
2.6. Generalizations of Precupanu s Inequality

2.6.1. Introduction. In 1976, T. Precupanu [29] obtained the
following result related to the Schwarz inequality in a real inner product
space (H; h, i) :
66
Theorem 23 (Precupanu , 1976). For any a H, x, y H\ {0} ,

we have the inequality:
kak kbk + ha, bi
2
hx, ai hx, bi hy, ai hy, bi
hx, ai hy, bi hx, yi
+
2
2
2
kxk
kyk
kxk2 kyk2
kak kbk + ha, bi
.
2
In the right-hand side or in the left-hand side of (2.100) we have equality if and only if there are , R such that
(2.100)
(2.101)
hy, bi
1
hx, ai
(a + b) .
2 x+
2 y =
2
kxk
kyk
Note for instance that [29], if y b, i.e., hy, bi = 0, then by (2.100)

one may deduce:
kak kbk + ha, bi
kak kbk + ha, bi
kxk2 hx, ai hx, bi
kxk2
2
2
for any a, b, x H, an inequality that has been obtained previously by
U. Richard [30]. The case of equality in the right-hand side or in the
left-hand side of (2.102) holds if and only if there are , R with
(2.102)
(2.103)
2 hx, ai x = (a + b) kxk2 .
For a = b, we may obtain from (2.100) the following inequality [29]

(2.104)
hx, ai2 hy, ai2

hx, ai hy, ai hx, yi
0
kak2 .
2 +
2 2
2
2
kxk
kyk
kxk kyk
This inequality implies [29]:

2
hx, yi
1 hx, ai
hy, ai
3
(2.105)
+
.
kxk kyk
2 kxk kak kyk kak
2
In [27], M.H. Moore pointed out the following reverse of the Schwarz
inequality
(2.106)
|hy, zi| kyk kzk ,
y, z H,
where some information about a third vector x is known:

Theorem 24 (Moore, 1973). Let (H; h, i) be an inner product
space over the real field R and x, y, z H such that:
(2.107)
|hx, yi| (1 ) kxk kyk ,
|hx, zi| (1 ) kxk kzk ,
67
where is a positive real number, reasonably small. Then

n
o
(2.108)
|hy, zi| max 1 2, 1 4, 0 kyk kzk .
Utilising Richards inequality (2.102) written in the following equivalent form:
hx, ai hx, bi
hx, ai hx, bi
(2.109) 2
kak kbk ha, bi 2
+ kak kbk
2
kxk
kxk2
for any a, b H and a H\ {0} , Precupanu has obtained the following
Moores type result:
Theorem 25 (Precupanu , 1976). Let (H; h, i) be a real inner
product space. If a, b, x H and 0 < 1 < 2 are such that:
(2.110)
1 kxk kak hx, ai 2 kxk kak ,

1 kxk kbk hx, bi 2 kxk kbk ,
then
(2.111)

221 1 kak kbk ha, bi 221 + 1 kak kbk .
Remark that the right inequality is always satisfied, since by Schwarzs

inequality, we have ha, bi kak kbk. The left inequality may be useful
when one assumes that 1 (0, 1]. In that case, from (2.111), we obtain

(2.112)
kak kbk 221 1 kak kbk ha, bi
provided 1 kxk kak hx, ai and 1 kxk kbk hx, bi , which is a refinement of Schwarzs inequality
kak kbk ha, bi .
In the complex case, apparently independent of Richard, M.L. Buzano
obtained in [3] the following inequality
kak kbk + |ha, bi|
(2.113)
|hx, ai hx, bi|
kxk2 ,
2
provided x, a, b are vectors in the complex inner product space (H; h, i) .
In the same paper [29], Precupanu , without mentioning Buzanos
name in relation to the inequality (2.113), observed that, on utilising
(2.113), one may obtain the following result of Moore type:
Theorem 26 (Precupanu , 1976). Let (H; h, i) be a (real or) complex inner product space. If x, a, b H are such that
(2.114)
|hx, ai| (1 ) kxk kak ,
|hx, bi| (1 ) kxk kbk ,
then
(2.115)

|ha, bi| 1 4 + 22 kak kbk .
68
Note that the above theorem is

for (0, 1], the
useful when,
i
2
2
quantity 1 4 + 2 > 0, i.e., 0, 1 2 .
Remark 26. When the space is real, the inequality (2.115) provides
a better lower bound for |ha, bi| than the second bound in Moores result
(2.108). However, it is not known if the first bound in (2.108) remains
valid for the case of complex spaces. From Moores original proof, apparently, the fact that the space (H; h, i) is real plays an essential role.
Before we point out some new results for orthonormal families of
vectors in real or complex inner product spaces, we state the following
result that complements the Moore type results outlined above for real
spaces [10]:
Theorem 27 (Dragomir, 2004). Let (H; h, i) be a real inner product space and a, b, x, y H\ {0} .
(i) If there exist 1 , 2 (0, 1] such that
hx, ai
1,
kxk kak
hy, ai
2
kyk kak
and 1 + 2 1, then
(2.116)
hx, yi
1
3
( 1 + 2 )2
kxk kyk
2
2
( 1) .
(ii) If there exist 1 (2 ) R such that

1 kak kbk
hx, ai hx, bi
( 2 kak kbk)
kxk2
and 1 1 0 (1 2 0) , then
(2.117)
[1 ] 21 1
ha, bi
( 22 + 1 [ 1]) .
kak kbk
The proof is obvious by the inequalities (2.105) and (2.109). We

omit the details.
2.6.2. Inequalities for Orthonormal Families. The following
result may be stated [10].
Theorem 28 (Dragomir, 2004). Let {ei }iI and {fj }jJ be two
finite families of orthonormal vectors in (H; h, i) . For any x, y
69
H\ {0} one has the inequality

X
X

(2.118)
hx, ei i hei , yi +
hx, fj i hfj , yi

iI
jJ

1
1

2
hx, ei i hfj , yi hei , fj i hx, yi kxk kyk .
2
2
iI,jJ
X
The case of equality holds in (2.118) if and only if there exists a K

such that
!
X
X
(2.119)
x y = 2
hx, ei i ei
hy, fj i fj .
iI
jJ

We know that, if u, v H, v 6= 0, then

2
2
2
2

hu,
vi
u
= kuk kvk |hu, vi|
(2.120)
v

kvk2
kvk2
showing that, in Schwarzs inequality
|hu, vi|2 kuk2 kvk2 ,
(2.121)
the case of equality, for v 6= 0, holds if and only if

hu, vi
(2.122)
u=
v,
kvk2
i.e. there exists a P
R such that u = v.
P
Now, let u := 2 iI hx, ei i ei x and v := 2 jJ hy, fj i fj y.
Observe that

2
*
+
X

X

kuk2 = 2
hx, ei i ei 4 Re
hx, ei i ei , x + kxk2

iI
iI
X
X
2
=4
|hx, ei i| 4
|hx, ei i|2 + kxk2 = kxk2 ,
iI
iI
and, similarly
kvk2 = kyk2 .
Also,
hu, vi = 4
hx, ei i hfj , yi hei , fj i + hx, yi
iI,jJ
X
iI
hx, ei i hei , yi 2
X
jJ
hx, fj i hfj , yi .
70
Therefore, by Schwarzs inequality (2.121) we deduce the desired inequality (2.118). By (2.122), the case of equality holds in (2.118) if
and only if there exists a K such that
!
X
X
2
hx, ei i ei x = 2
hy, fj i fj y ,
iI
jJ
which is equivalent to (2.119).

Remark 27. If in (2.119) we choose x = y, then we get the inequality:

X
X

(2.123)
|hx, ei i|2 +
|hx, fj i|2

iI
jJ

1
X
1

2
hx, ei i hfj , xi hei , fj i kxk2 kxk2
2
2
iI,jJ
for any x H.
If in the above theorem we assume that I = J and fi = ei , i I,
then we get from (2.118) the Schwarz inequality |hx, yi| kxk kyk .
If I J = , I J = K, gk = ek , k I, gk = fk , k J and
{gk }kK is orthonormal, then from (2.118) we get:

X
1
1

(2.124)
hx, gk i hgk , yi hx, yi kxk kyk ,
x, y H

2
2
kK
which has been obtained earlier by the author in [16].
e
If I and J reduce to one element, namely e1 = kek
, f1 = kff k with
e, f 6= 0, then from (2.118) we get

hx, ei he, yi hx, f i hf, yi

hx, ei hf, yi he, f i 1

(2.125)
+
hx,
yi

2
kek2
kf k2
kek2 kf k2
1
kxk kyk ,
x, y H
2
which is the corresponding complex version of Precupanu s inequality
(2.100).
If in (2.125) we assume that x = y, then we get

|hx, ei|2 |hx, f i|2

hx, ei hf, ei he, f i 1

2
(2.126)
+
kxk

kek2

2
kf k2
kek2 kf k2
1
kxk2 .
2
71
The following corollary may be stated [10]:

Corollary 14 (Dragomir, 2004). With the assumptions of Theorem 28, we have:

X
X

(2.127)
hx, ei i hei , yi +
hx, fj i hfj , yi

iI
jJ

X

2
hx, ei i hfj , yi hei , fj i

iI,jJ

X
X
1

hx, ei i hei , yi +
hx, fj i hfj , yi
|hx, yi| +

2
iI
jJ

X
1

2
hx, ei i hfj , yi hei , fj i |hx, yi|

2
iI,jJ
1
[|hx, yi| + kxk kyk] .
2
Proof. The first inequality follows by the triangle inequality for

the modulus. The second inequality follows by (2.118) on adding the
quantity 21 |hx, yi| on both sides.
Remark 28.
(1) If we choose in (2.127), x = y, then we get:

X
X

(2.128)
|hx, ei i|2 +
|hx, fj i|2

iI
jJ

X

2
hx, ei i hfj , xi hei , fj i

iI,jJ

X
X

|hx, ei i|2 +
|hx, fj i|2

iI
jJ

1
X
1

2
hx, ei i hfj , xi hei , fj i kxk2 + kxk2
2
2
iI,jJ
kxk .
We observe that (2.128) will generate Bessels inequality if
{ei }iI , {fj }jJ are disjoint parts of a larger orthonormal family.
72
(2) From (2.125) one can obtain:

hx, ei he, yi hx, f i hf, yi
hx, ei hf, yi he, f i

(2.129)
+
2

kek2
kf k2
kek2 kf k2
1
[kxk kyk + |hx, yi|]
2
and in particular

|hx, ei|2 |hx, f i|2
hx, ei hf, ei he, f i

2
+
2
(2.130)
kxk ,

2
2
2
2

kek
kf k
kek kf k
for any x, y H.
The case of real inner products will provide a natural genearlization
for Precupanu s inequality (2.100) [10]:
Corollary 15 (Dragomir, 2004). Let (H; h, i) be a real inner
product space and {ei }iI , {fj }jJ two finite families of orthonormal
vectors in (H; h, i) . For any x, y H\ {0} one has the double inequality:
X
X
1
(2.131)
[|hx, yi| kxk kyk]
hx, ei i hy, ei i +
hx, fj i hy, fj i
2
iI
jJ
X
2
hx, ei i hy, fj i hei , fj i
iI,jJ
1
[kxk kyk + |hx, yi|] .
2
In particular, we have
X
X
X
(2.132) 0
hx, ei i2 +
hx, fj i2 2
hx, ei i hx, fj i hei , fj i
iI
jJ
iI,jJ
kxk ,
for any x H.
Remark 29. Similar particular inequalities to those incorporated
in (2.124) (2.130) may be stated, but we omit them.
2.6.3. Refinements of Kurepas Inequality. The following result may be stated [10].
Theorem 29 (Dragomir, 2004). Let (H; h, i) be a real inner product space and {ei }iI , {fj }jJ two finite families in H. If (HC ; h, iC )
73
is the complexification of (H; h, i) , then for any w HC , we have the

inequalities

X

X
X

(2.133)
hw, ei i2C +
hw, fj i2C 2
hw, ei iC hw, fj iC hei , fj i

iI
jJ
iI,jJ

X
X
1

C| +
hw, ei i2C +
hw, fj i2C
|hw, wi

2
iI
jJ

X
1

2
hw, ei iC hw, fj iC hei , fj i hw, wi
C

2
iI,jJ

1
kwk2C + |hw, wi
C | kwk2C .
2
Proof. Define gj HC , gj := (ej , 0) , j I. For any k, j I we
have
hgk , gj iC = h(ek , 0) , (ej , 0)iC = hek , ej i = kj ,
therefore {gj }jI is an orthonormal family in (HC ; h, iC ) .
If we apply Corollary 14 for (HC ; h, iC ) , x = w, y = w,
we may
write:

X
X

(2.134)
hw, ei iC hei , wi
C+
hw, fj i hfj , wi

jJ
iI

X

2
hw, ei iC hfj , wiC hei , fj i

iI,jJ

X
X
1

kwkC kwk
C+
hw, ei iC hei , wi
C+
hw, fj i hfj , wi

2
iI
jJ

X
1

C
2
hw, ei iC hfj , wiC hei , fj i hw, wi

2
iI,jJ
1
[|hw, wi
C | + kwkC kwk
C] .
2
However, for w := (x, y) HC , we have w = (x, y) and

hej , wi
C = h(ej , 0) , (x, y)iC = hej , xi + i hej , yi
and
hw, ej iC = h(x, y) , (ej , 0)iC = hx, ej i + i hej , yi
74
showing that hej , wi

= hw, ej i for any j I. A similar relation is true
for fj and since
1
kwkC = kwk
C = kxk2 + kyk2 2 ,
hence from (2.134) we deduce the desired inequality (2.133).
Remark 30. It is obvious that, if one family, say {fj }jJ is empty,
P
then, on observing that all sums jJ should be zero, from (2.133) one
would get [16]

X

(2.135)
hw, ei i2C

iI

X
1
1

2
C| +
hw, ei iC hw, wi
C
|hw, wi

2
2
iI

1
kwk2C + |hw, wi
C | kwk2C .
2
If in (2.135) one assumes that the family {ei }iI contains only one
a
element e = kak
, a 6= 0, then by selecting w = z, one would deduce
(2.71), which is a refinement for Kurepas inequality.
2.7. Some New Refinements of the Schwarz Inequality
2.7.1. Refinements. The following result holds [12].
space over the real or complex number field K and r1 , r2 > 0. If x, y H
are with the property that
(2.136)
kx yk r2 r1 |kxk kyk| ,
then we have the following refinement of Schwarzs inequality

1 2
(2.137)
kxk kyk Re hx, yi
r2 r12 ( 0) .
2
The constant 21 is best possible in the sense that it cannot be replaced
by a larger quantity.
Proof. From the first inequality in (2.136) we have
(2.138)
kxk2 + kyk2 r22 + 2 Re hx, yi .
Subtracting in (2.138) the quantity 2 kxk kyk , we get

(2.139)
(kxk kyk)2 r22 2 (kxk kyk Re hx, yi) .
2.7. SOME NEW REFINEMENTS OF THE SCHWARZ INEQUALITY
75
Since, by the second inequality in (2.136) we have

r12 (kxk kyk)2 ,
(2.140)
hence from (2.139) and (2.140) we deduce the desired inequality (2.137).
To prove the sharpness of the constant 12 in (2.137), let us assume
that there is a constant C > 0 such that

(2.141)
kxk kyk Re hx, yi C r22 r12 ,
provided that x and y satisfy (2.136).
Let e H with kek = 1 and for r2 > r1 > 0, define
r 1 r2
r2 + r 1
(2.142)
x=
e and y =
e.
2
2
Then
kx yk = r2 and |kxk kyk| = r1 ,
showing that the condition (2.136) is fulfilled with equality.
If we replace x and y as defined in (2.142) into the inequality (2.141),
then we get

r22 r12
C r22 r12 ,
2
which implies that C 21 , and the theorem is completely proved.
The following corollary holds.
Corollary 16. With the assumptions of Theorem 30, we have the
inequality:
q
2
2
(2.143)
kxk + kyk
kx + yk
r22 r12 .
2
2
Proof. We have, by (2.137), that
(kxk + kyk)2 kx + yk2 = 2 (kxk kyk Re hx, yi) r22 r12 0
which gives
(2.144)
(kxk + kyk) kx + yk +
2
q
r22 r12
By making use of the elementary inequality

2 2 + 2 ( + )2 ,
, 0;
we get
(2.145)
kx + yk +
q
2
r22
r12

kx + yk +
2
r22
r12
Utilising (2.144) and (2.145), we deduce the desired inequality (2.143).
76
If (H; h, i) is a Hilbert space and {ei }iI is an orthornormal family

in H, i.e., we recall that hei , ej i = ij for any i, j I, where ij is
Kroneckers delta, then we have the following inequality which is well
known in the literature as Bessels inequality
X
(2.146)
|hx, ei i|2 kxk2 for each x H.
iI
Here, the meaning of the sum is

(
)
X
X
|hx, ei i|2 = sup
|hx, ei i|2 , F is a finite part of I .
F I
iI
iF
The following result providing a refinement of the Bessel inequality

(2.146) holds [12].
Theorem 31 (Dragomir, 2004). Let (H; h, i) be a Hilbert space
and {ei }iI an orthonormal family in H. If x H, x 6= 0, and r2 , r1 > 0
are such that:

X

(2.147) x
hx, ei i ei

iI
! 21
X
|hx, ei i|2
r2 r1 kxk
( 0) ,
iI
then we have the inequality

! 21
(2.148)
kxk
|hx, ei i|2
iI
1
r2 r12
P 2
1 ( 0) .
2 2
2
iI |hx, ei i|
1
2
is best possible.
P
Proof. Consider y := iI hx, ei i ei . Obviously, since H is a Hilbert
space, y H. We also note that

v

2 s
X
u
X

X

u

t
kyk =
hx, ei i ei =
|hx, ei i|2 ,
hx, ei i ei =

The constant
iI
iI
and thus (2.147) is in fact (2.136) of Theorem 30.
iI
77
Since
! 12
kxk kyk Re hx, yi = kxk
*
Re x,
|hx, ei i|
+
X
iI
hx, ei i ei
iI
! 12
! 21
=
|hx, ei i|2
kxk
|hx, ei i|2
iI
iI
hence, by (2.137), we deduce the desired result (2.148).

We will prove the sharpness of the constant for the case of one
element, i.e., I = {1} , e1 = e H, kek = 1. For this, assume that
there exists a constant D > 0 such that
r2 r12
(2.149)
kxk |hx, ei| D 2
|hx, ei|
provided x H\ {0} satisfies the condition
(2.150)
kx hx, ei ek r2 r1 kxk |hx, ei| .
Assume that x = e + f with e, f H, kek = kf k = 1 and e f. We

wish to see if there exists positive numbers , such that
(2.151)
kx hx, ei ek = r2 > r1 = kxk |hx, ei| .
Since (for , > 0)

kx hx, ei ek =
and
q
kxk |hx, ei| = 2 + 2
hence, by (2.151), we get = r2 and

q
2 + r22 = r1
giving
2 + r22 = 2 + 2r1 + r12
from where we get
r22 r12
> 0.
2r1
With these values for and , we have
=
kxk |hx, ei| = r1 ,
|hx, ei| =
and thus, from (2.149), we deduce

r1 D
r22 r12
r22 r12
2r1
r22 r12
2r1
78
giving D 12 . This proves the theorem.

The following corollary is obvious.
Corollary 17. Let x, y H with hx, yi =
6 0 and r2 r1 > 0 such
that

hx, yi

(2.152)
kyk x kyk y r2 kyk r1 kyk
kxk kyk |hx, yi| ( 0) .
Then we have the following refinement of the Schwarzs inequality:
kxk kyk |hx, yi|
(2.153)
The constant
1
2
kyk2
1 2
r2 r12
( 0) .
2
|hx, yi|
is best possible.
The following lemma holds [12].

Lemma 3 (Dragomir, 2004). Let (H; h, i) be an inner product space
and R 1. For x, y H, the subsequent statements are equivalent:
(i) The following refinement of the triangle inequality holds:
(2.154)
kxk + kyk R kx + yk ;
(ii) The following refinement of the Schwarz inequality holds:

1 2
R 1 kx + yk2 .
(2.155)
kxk kyk Re hx, yi
2
Proof. Taking the square in (2.154), we have

(2.156) 2 kxk kyk R2 1 kxk2 + 2R2 Re hx, yi + R2 1 kyk2 .
Subtracting from both sides of (2.156) the quantity 2 Re hx, yi , we obtain

2 (kxk kyk Re hx, yi) R2 1 kxk2 + 2 Re hx, yi + kyk2

= R2 1 kx + yk2 ,
which is clearly equivalent to (2.155).
By the use of the above lemma, we may now state the following
theorem concerning another refinement of the Schwarz inequality [12].
space over the real or complex number field and R 1, r 0. If
x, y H are such that
1
(2.157)
(kxk + kyk) kx + yk r,
R
79
then we have the following refinement of the Schwarz inequality

1 2
R 1 r2 .
(2.158)
kxk kyk Re hx, yi
2
1
Proof. The inequality (2.158) follows easily from Lemma 3. We
need only prove that 12 is the best possible constant in (2.158).
Assume that there exists a C > 0 such that

(2.159)
kxk kyk Re hx, yi C R2 1 r2
provided x, y, R and r satisfy (2.157).
Consider r = 1, R > 1 and choose x = 1R
e, y = 1+R
e with e H,
2
2
kek = 1. Then
kxk + kyk
x + y = e,
=1
R
and thus (2.157) holds with equality on both sides.
From (2.159), for the above choices, we have 12 (R2 1) C (R2 1) ,
which shows that C 12 .
Finally, the following result also holds [12].
space over the real or complex number field K and r (0, 1]. For x, y
H, the following statements are equivalent:
(i) We have the inequality
(2.160)
|kxk kyk| r kx yk ;
(ii) We have the following refinement of the Schwarz inequality

1
(2.161)
kxk kyk Re hx, yi
1 r2 kx yk2 .
2
1
The constant 2 in (2.161) is best possible.
Proof. Taking the square in (2.160), we have
kxk2 2 kxk kyk + kyk2 r2 kxk2 2 Re hx, yi + kyk2
which is clearly equivalent to

1 r2 kxk2 2 Re hx, yi + kyk2 2 (kxk kyk Re hx, yi)
or with (2.161).
Now, assume that (2.161) holds with a constant E > 0, i.e.,

(2.162)
kxk kyk Re hx, yi E 1 r2 kx yk2 ,
provided (2.160) holds.
80
Define x =
r+1
e,
2
y=
r1
e
2
with e H, kek = 1. Then
|kxk kyk| = r,
kx yk = 1
showing that (2.160) holds with equality.

If we replace x and y in (2.162), then we get E (1 r2 ) 12 (1 r2 ) ,
implying that E 12 .
2.7.2. Discrete Inequalities. Assume that (K; (, )) is a Hilbert
space over the
Preal or complex number field . Assume also that pi 0,
i H with i=1 pi = 1 and define
(
)
X

`2p (K) := x := (xi ) xi K, i N and
pi kxi k2 < .
iN
i=1
It is well known that `2p (K) endowed with the inner product h, ip
defined by
X
hx, yip :=
pi (xi , yi )
i=1
and generating the norm

kxkp :=
! 21
pi kxi k2
i=1
is a Hilbert space over K.

We may state the following discrete inequality improving the CauchyBunyakovsky-Schwarz classical result [12].
Proposition
P 16. Let (K; (, )) be a Hilbert2 space and pi 0
(i N) with i=1 pi = 1. Assume that x, y `p (K) and r1 , r2 > 0
satisfy the condition
kxi yi k r2 r1 |kxi k kyi k|
(2.163)
for each i N. Then we have the following refinement of the CauchyBunyakovsky-Schwarz inequality
(2.164)
X
i=1
pi kxi k
X
2
! 12
pi kyi k2
i=1
pi Re (xi , yi )
i=1
The constant
1
2
is best possible.

1 2
r2 r12 0.
2
81
Proof. From the condition (2.163) we simply deduce
X
X
2
2
2
(2.165)
pi kxi yi k r2 r1
pi (kxi k kyi k)2
i=1
i=1
! 12
pi kxi k2
i=1
! 12 2
pi kyi k2
i=1
In terms of the norm kkp , the inequality (2.165) may be written as

(2.166)
kx ykp r2 r1 kxkp kykp .

Utilising Theorem 30 for the Hilbert space `2p (K) , h, ip , we deduce
the desired inequality (2.164).
For n = 1 (p1 = 1) , the inequality (2.164) reduces to (2.137) for
which we have shown that 21 is the best possible constant.
By the use of Corollary 16, we may state the following result as
well.
Corollary 18. With the assumptions of Proposition 16, we have
the inequality
! 12
! 12
X
X
(2.167)
pi kxi k2
+
pi kyi k2
i=1
i=1
! 12
pi kxi + yi k
i=1
q
2
r22 r12 .
2
The following proposition also holds [12].

Proposition
P 17. Let (K; (, )) be a Hilbert 2space and pi 0
(i N) with
i=1 pi = 1. Assume that x, y `p (K) and R 1,
r 0 satisfy the condition
1
(2.168)
(kxi k + kyi k) kxi + yi k r
R
for each i N. Then we have the following refinement of the Schwarz
inequality
! 12
X
X
X
2
2
(2.169)
pi kxi k
pi kyi k
pi Re (xi , yi )
i=1
i=1
i=1

1 2
R 1 r2 .
2
82

Proof. By (2.168) we deduce
"
# 12
1 X
pi (kxi k + kyi k)2
(2.170)
R i=1
! 12
pi kxi + yi k2
r.
i=1
By the classical Minkowsky inequality for nonnegative numbers, we

have
! 12
! 12
X
X
(2.171)
pi kxi k2
+
pi kyi k2
i=1
i=1
"
# 12
pi (kxi k + kyi k)2
i=1
and thus, by utilising (2.170) and (2.171), we may state in terms of

kkp the following inequality

1
(2.172)
kxkp + kykp kx + ykp r.
R
Employing Theorem 32 for the Hilbert space `2p (K) and the inequality
(2.172), we deduce the desired result (2.169).
Since, for p = 1, n = 1, (2.169) reduced to (2.158) for which we
have shown that 21 is the best constant, we conclude that 12 is the best
constant in (2.169) as well.
Finally, we may state and prove the following result [12] incorporated in
Proposition
P 18. Let (K; (, )) be a Hilbert2 space and pi 0
(i N) with i=1 pi = 1. Assume that x, y `p (K) and r (0, 1]
such that
|kxi k kyi k| r kxi yi k for each i N,
(2.173)
holds true. Then we have the following refinement of the Schwarz inequality
! 21
X
X
X
pi Re (xi , yi )
(2.174)
pi kxi k2
pi kyi k2
i=1
i=1
i=1
X
1
1 r2
pi kxi yi k2 .
2
i=1
The constant
1
2
83
is best possible in (2.174).
Proof. From (2.173) we have

# 12
"
# 12
"
X
X
pi (kxi k kyi k)2 r
pi kxi yi k2 .
i=1
i=1
Utilising the following elementary result

! 21
! 12

X
X

pi kxi k2
pi kyi k2

i=1

i=1
! 12
pi (kxi k kyi k)2
i=1
we may state that

kxkp kykp r kx ykp .
Now, by making use of Theorem 33, we deduce the desired inequality
(2.174) and the fact that 21 is the best possible constant. We omit the
details.
2.7.3. Integral Inequalities. Assume that (K; (, )) is a Hilbert
space over the real or complex number field K. If : [a, b] R [0, )
Rb
is a Lebesgue integrable function with a (t) dt = 1, then we may
consider the space L2 ([a, b] ; K) of all functions f : [a, b] K, that
Rb
are Bochner measurable and a (t) kf (t)k2 dt < . It is known that
L2 ([a, b] ; K) endowed with the inner product h, i defined by
Z b
hf, gi :=
(t) (f (t) , g (t)) dt
a

Z
kf k :=
21
(t) kf (t)k dt
2

Now we may state and prove the first refinement of the CauchyBunyakovsky-Schwarz integral inequality [12].
Proposition 19. Assume that f, g L2 ([a, b] ; K) and r2 , r1 > 0
satisfy the condition
(2.175)
kf (t) g (t)k r2 r1 |kf (t)k kg (t)k|
84
for a.e. t [a, b] . Then we have the inequality

b
Z
(t) kf (t)k dt
(2.176)
a
12
(t) kg (t)k dt
2
(t) Re (f (t) , g (t)) dt

a
1
2
The constant

1 2
r2 r12 ( 0) .
2
Proof. Integrating (2.175), we get

b
Z
21
(t) (kf (t) g (t)k) dt
2
(2.177)
a
Z
r2 r 1
12
(t) (kf (t)k kg (t)k)2 dt .
Utilising the obvious fact

Z
(2.178)
a
12
(t) (kf (t)k kg (t)k) dt
Z
12
12 Z b

b

(t) kg (t)k2 dt ,

(t) kf (t)k2 dt
a

a
2
we can state the following inequality in terms of the kk norm:

(2.179)
kf gk r2 r1 kf k kgk .
Employing Theorem 30 for the Hilbert space L2 ([a, b] ; K) , we deduce
the desired inequality (2.176).
To prove the sharpness of 21 in (2.176), we choose a = 0, b = 1,
f (t) = 1, t [0, 1] and f (t) = x, g (t) = y, t [a, b] , x, y K. Then
(2.176) becomes

1 2
kxk kyk Re hx, yi
r2 r12
2
provided
kx yk r2 r1 |kxk kyk| ,
which, by Theorem 30 has the quantity
stant.
The following corollary holds.
1
2
as the best possible con-
85
Corollary 19. With the assumptions of Proposition 19, we have

the inequality
Z
12
12 Z b
2
(t) kg (t)k dt
(t) kf (t)k dt
+
2
(2.180)
q
Z b
12
2
2
2
(t) kf (t) + g (t)k dt
r22 r12 .
2
2
a
The following two refinements of the Cauchy-Bunyakovsky-Schwarz

(CBS) integral inequality also hold.
Proposition 20. If f, g L2 ([a, b] ; K) and R 1, r 0 satisfy
the condition
1
(2.181)
(kf (t)k + kg (t)k) kf (t) + g (t)k r
R
for a.e. t [a, b] , then we have the inequality
Z
(t) kf (t)k dt
(2.182)
a
12
(t) kg (t)k dt
2
a
b
(t) Re (f (t) , g (t)) dt

a
The constant
1
2

1 2
R 1 r2 .
2
The proof follows by Theorem 32 and we omit the details.

Proposition 21. If f, g L2 ([a, b] ; K) and (0, 1] satisfy the
condition
|kf (t)k kg (t)k| kf (t) g (t)k
(2.183)

Z
b
2
(t) kf (t)k dt
(2.184)
a
12
(t) kg (t)k dt
2
The constant
1
2
(t) Re (f (t) , g (t)) dt

Z
b
1
2
1
(t) kf (t) g (t)k2 dt.
2
a
The proof follows by Theorem 33 and we omit the details.
86
2.7.4. Refinements of the Heisenberg Inequality. It is well

known that if (H; h, i) is a real or complex Hilbert space and f :
[a, b] R H is an absolutely continuous vector-valued function, then
f is differentiable almost everywhere on [a, b] , the derivative f 0 : [a, b]
H is Bochner integrable on [a, b] and
Z t
(2.185)
f (t) =
f 0 (s) ds
for any t [a, b] .
a
The following theorem provides a version of the Heisenberg inequalities in the general setting of Hilbert spaces [12].
Theorem 34 (Dragomir, 2004). Let : [a, b] H be an absolutely
continuous function with the property that b k (b)k2 = a k (a)k2 .
Then we have the inequality:
Z b
2
Z b
Z b
2
2
2
2
(2.186)
k (t)k dt 4
t k (t)k dt
k0 (t)k dt.
a

by a smaller constant.
Proof. Integrating by parts, we have successively
Z b
(2.187)
k (t)k2 dt
a
b Z b

d
2
= t k (t)k
t
k (t)k2 dt
dt
a
a
Z b
d
2
2
= b k (b)k a k (a)k
t h (t) , (t)i dt
dt
a
Z b
=
t [h0 (t) , (t)i + h (t) , 0 (t)i] dt
a
Z b
= 2
t Re h0 (t) , (t)i dt
a
Z b
=2
Re h0 (t) , (t) (t)i dt.
a
If we apply the Cauchy-Bunyakovsky-Schwarz integral inequality

Z b
12
Z b
Z b
2
2
Re hg (t) , h (t)i dt
kg (t)k dt
kh (t)k dt
a
a
0
for g (t) = (t) , h (t) = t (t) , t [a, b] , then we deduce the desired
inequality (2.176).
87
The fact that 4 is the best constant in (2.176) follows from the fact
that in the (CBS) inequality, the case of equality holds iff g (t) = h (t)
for a.e. t [a, b] and a given scalar in K. We omit the details.
For details on the classical Heisenberg inequality, see, for instance,
[23].
Utilising Proposition 19, we can state the following refinement [12]
of the Heisenberg inequality obtained above in (2.186):
Proposition 22. Assume that : [a, b] H is as in the hypothesis
of Theorem 34. In addition, if there exists r2 , r1 > 0 so that
k0 (t) + t (t)k r2 r1 |k0 (t)k |t| k (t)k|
Z
b
2
t k (t)k dt
a
12
Z
1 b
k (t)k dt
k (t)k2 dt
2 a

1
(b a) r22 r12 ( 0) .
2
0
The proof follows by Proposition 19 on choosing f (t) = 0 (t) , g (t) =

1
t (t) and (t) = ba
, t [a, b] .
On utilising the Proposition 20 for the same choices of f, g and ,
we may state the following results as well [12]:
Proposition 23. Assume that : [a, b] H is as in the hypothesis
of Theorem 34. In addition, if there exists R 1 and r > 0 so that
1
(k0 (t)k + |t| k (t)k) k0 (t) t (t)k r
R
Z
a
t2 k (t)k2 dt
21
Z
1 b
2
0
k (t)k2 dt
k (t)k dt
2 a

1
(b a) R2 1 r2 ( 0) .
2
Finally, we can state

Proposition 24. Let : [a, b] H be as in the hypothesis of
Theorem 34. In addition, if there exists (0, 1] so that
|k0 (t)k |t| k (t)k| k0 (t) + t (t)k
88

Z
b
2
t2 k (t)k dt
Z
a
21
Z
1 b
0
k (t)k2 dt
k (t)k dt
2 a
Z b

1
2
1 2
k0 (t) + t (t)k dt ( 0) .
2
a
2
This follows by Proposition 21 and we omit the details.

2.8. More Schwarz Related Inequalities
2.8.1. Introduction. In practice, one may need reverses of the
Schwarz inequality, namely, upper bounds for the quantities
kxk kyk Re hx, yi ,
kxk2 kyk2 (Re hx, yi)2
and
kxk kyk
Re hx, yi
or the corresponding expressions where Re hx, yi is replaced by either
|Re hx, yi| or |hx, yi| , under suitable assumptions for the vectors x, y in
an inner product space (H; h, i) over the real or complex number field
K.
In this class of results, we mention the following recent reverses of
the Schwarz inequality due to the present author, that can be found,
for instance, in the survey work [4], where more specific references are
provided:
space over K (K = C, R) . If a, A K and x, y H are such that either
Re hAy x, x ayi 0,
(2.188)
or, equivalently,

1
A
+
a
x
(2.189)
y

2 |A a| kyk ,
2
then the following reverse for the quadratic form of the Schwarz inequality
(2.190)
(0 ) kxk2 kyk2 |hx, yi|2

2
2

14 |A a|2 kyk4 A+a
kyk
hx,
yi
2
1
|A a|2 kyk4 kyk2 Re hAy x, x ayi
4
1
|A a|2 kyk4
4
2.8. MORE SCHWARZ RELATED INEQUALITIES
89
holds.
If in addition, we have Re (A
a) > 0, then

hx, yi
1 Re A + a
|A + a|
1
p
(2.191) kxk kyk
p
|hx, yi| ,
2
2
Re (A
a)
Re (A
a)
and
(2.192)
1 |A a|2
(0 ) kxk kyk |hx, yi|
|hx, yi|2 .
4 Re (A
a)
2
Also, if (2.188) or (2.189) are valid and A 6= a, then we have the

reverse for the simple form of Schwarz inequality

+a

A
(2.193) (0 ) kxk kyk |hx, yi| kxk kyk Re

hx, yi
|A + a|

A+a
1 |A a|2
hx, yi
kyk2 .
kxk kyk Re
|A + a|
4 |A + a|
The multiplicative constants 14 and 21 above are best possible in the sense
that they cannot be replaced by a smaller quantity.
For some classical results related to Schwarz inequality, see [3], [21],
[28], [29], [30] and the references therein.
The main aim of the present section is to point out other results in
connection with both the quadratic and simple forms of the Schwarz
inequality. As applications, some reverse results for the generalised
triangle inequality, i.e., upper bounds for the quantity

n
n
X

X

(0 )
kxi k
xi

i=1
i=1
under various assumptions for the vectors xi H, i {1, . . . , n} , are

established.
2.8.2. Refinements and Reverses. The following result holds
[7].
Proposition 25. Let (H; h, i) be an inner product space over the
real or complex number field K. The subsequent statements are equivalent.
(i) The following inequality holds

x
y

(2.194)
kxk kyk () r;
90
(ii) The following reverse (improvement) of Schwarzs inequality

holds
1
(2.195)
kxk kyk Re hx, yi () r2 kxk kyk .
2
1
The constant 2 is best possible in (2.195) in the sense that it
cannot be replaced by a larger (smaller) quantity.
Remark 31. Since
kkyk x kxk yk = kkyk (x y) + (kyk kxk) yk
kyk kx yk + |kyk kxk| kyk
2 kyk kx yk
hence a sufficient condition for (2.194) to hold is
r
(2.196)
kx yk kxk .
2
Remark 32. Utilising the Dunkl-Williams inequality [20]

a

1
b

(2.197)
ka bk (kak + kbk)
kak kbk , a, b H\ {0}

2
with equality if and only if either kak = kbk or kak + kbk = ka bk ,
we can state the following inequality

2
kxk kyk Re hx, yi
kx yk
(2.198)
2
, x, y H\ {0} .
kxk kyk
kxk + kyk
Obviously, if x, y H\ {0} are such that
(2.199)
kx yk (kxk + kyk) ,
with (0, 1], then one has the following reverse of the Schwarz
inequality
(2.200)
kxk kyk Re hx, yi 2 2 kxk kyk
that is similar to (2.195).

The following result may be stated as well [7].
Proposition 26. If x, y H\ {0} and > 0 are such that

x
y

(2.201)
kyk kxk ,
then we have the following reverse of Schwarzs inequality
(2.202)
(0 ) kxk kyk |hx, yi| kxk kyk Re hx, yi

1
2 kxk kyk .
2
91
The case of equality holds in the last inequality in (2.202) if and only
if
kxk = kyk
(2.203)
The constant
1
2
and
kx yk = .
in (2.202) cannot be replaced by a smaller quantity.
Proof. Taking the square in (2.201), we get

kxk2 2 Re hx, yi kyk2
+
2 .
kxk kyk
kyk2
kxk2
(2.204)
Since, obviously
kxk2 kyk2
2
+
kyk2 kxk2
(2.205)
with equality iff kxk = kyk , hence by (2.204) we deduce the second
inequality in (2.202).
The case of equality and the best constant are obvious and we omit
the details.
Remark 33. In [24], Hile obtained the following inequality
(2.206)
kxkv+1 kykv+1
kkxk x kyk yk
kx yk
kxk kyk
v
provided v > 0 and kxk =

6 kyk .
If in (2.206) we choose v = 1 and take the square, then we get
(2.207) kxk4 2 kxk kyk Re hx, yi + kyk4 (kxk + kyk)2 kx yk2 .
Since,
kxk4 + kyk4 2 kxk2 kyk2 ,
hence, by (2.207) we deduce
(2.208)
1 (kxk + kyk)2 kx yk2

(0 ) kxk kyk Re hx, yi
,
2
kxk kyk
provided x, y H\ {0} .
The following inequality is due to Goldstein, Ryff and Clarke [22,
p. 309]:
Re hx, yi
(2.209) kxk2r + kyk2r 2 kxkr kykr
kxk kyk
2r2
kx yk2 if r 1
r2 kxk
kyk2r2 kx yk2
if r < 1
92
provided r R and x, y H with kxk kyk .

Utilising (2.209) we may state the following proposition containing
a different reverse of the Schwarz inequality in inner product spaces
[7].
Proposition 27. Let (H; h, i) be an inner product space over the
real or complex number field K. If x, y H\ {0} and kxk kyk , then
we have
0 kxk kyk |hx, yi| kxk kyk Re hx, yi
r1
1 2 kxk
r
kx yk2 if r 1,
2
kyk

1 kxk 1r kx yk2
if r < 1.
(2.210)
kyk
Proof. It follows from (2.209), on dividing by kxkr kykr , that

r
r
kxk
kyk
Re hx, yi
(2.211)
+
2
kyk
kxk
kxk kyk
r2
2
2 kxk
r kykr kx yk if r 1,
kykr2 kx yk2
if r < 1.
kxkr
Since
kxk
kyk
hence, by (2.211) one has
22
r

+
kyk
kxk
r
2,
r2
2
2 kxk
r kykr kx yk if r 1,
Re hx, yi
kxk kyk
kykr2
kxkr
kx yk2
if r < 1.
Dividing this inequality by 2 and multiplying with kxk kyk , we deduce

the desired result in (2.210).
Another result providing a different additive reverse (refinement) of
the Schwarz inequality may be stated [7].
Proposition 28. Let x, y H with y 6= 0 and r > 0. The subsequent statements are equivalent:
(i) The following inequality holds:

hx,
yi
x
() r;
y
(2.212)

kyk2
93
(ii) The following reverse (refinement) of the quadratic Schwarz

inequality holds:
(2.213)
kxk2 kyk2 |hx, yi|2 () r2 kyk2 .
The proof is obvious on taking the square in (2.212) and performing

the calculation.
Remark 34. Since

kyk2 x hx, yi y = kyk2 (x y) hx y, yi y
kyk2 kx yk + |hx y, yi| kyk
2 kx yk kyk2 ,
hence a sufficient condition for the inequality (2.212) to hold is that
r
(2.214)
kx yk .
2
The following proposition may give a complementary approach [7]:
Proposition 29. Let x, y H with hx, yi =
6 0 and > 0. If

x hx, yi y ,
(2.215)

|hx, yi|
then
1
(0 ) kxk kyk |hx, yi| 2 .
2
1
The multiplicative constant 2 is best possible in (2.216).
(2.216)
The proof is similar to the ones outlined above and we omit it.
For the case of complex inner product spaces, we may state the
following result [7].
Proposition 30. Let (H; h, i) be a complex inner product space
and C a given complex number with Re , Im > 0. If x, y H
are such that

Im

(2.217)
x Re y r,
1 Re 2

r .
2 Im
The equality holds in the second inequality in (2.218) if and only if the
case of equality holds in (2.217) and Re kxk = Im kyk .
(2.218)
94
Proof. Observe that the condition (2.217) is equivalent to

(2.219) [Re ]2 kxk2 + [Im ]2 kyk2 2 Re Im Re hx, yi + [Re ]2 r2 .
On the other hand, on utilising the elementary inequality
(2.220)
2 Re Im kxk kyk [Re ]2 kxk2 + [Im ]2 kyk2 ,
with equality if and only if Re kxk = Im kyk , we deduce from

(2.219) that
(2.221)
2 Re Im kxk kyk 2 Re Im Re hx, yi + r2 [Re ]2
giving the desired inequality (2.218).

The case of equality follows from the above and we omit the details.
The following different reverse for the Schwarz inequality that holds
for both real and complex inner product spaces may be stated as well
[7].
space over K, K = C, R. If K\ {0} , then
2

(2.222)
0 kxk kyk |hx, yi| kxk kyk Re
hx, yi
||2
1 [|Re | kx yk + |Im | kx + yk]2

1
I 2,
2
2
2
||
where
(2.223)
max {|Re | , |Im |} (kx yk + kx + yk) ;
1
1
I :=
(|Re |p + |Im |p ) p (kx ykq + kx + ykq ) q ,
p > 1, p1 + 1q = 1;
max {kx yk , kx + yk} (|Re | + |Im |) .
Proof. Observe, for K\ {0} , that

kx
yk2 = ||2 kxk2 2 Re hx,
yi + ||2 kyk2

= ||2 kxk2 + kyk2 2 Re 2 hx, yi .
Since kxk2 + kyk2 2 kxk kyk , hence

2
2
2
(2.224)
kx
yk 2 || kxk kyk Re
hx, yi .
||2
95
On the other hand, we have

(2.225)
kx
yk = k(Re + i Im ) x (Re i Im ) yk
= kRe (x y) + i Im (x + y)k
|Re | kx yk + |Im | kx + yk .
Utilising (2.224) and (2.225) we deduce the third inequality in (2.222).

For the last inequality we use the following elementary inequality
max {, } (a + b)
(2.226) a + b
1
1
p
( + p ) p (aq + bq ) q , p > 1, p1 + 1q = 1,
provided , , a, b 0.
The following result may be stated [7].
Proposition 31. Let (H; h, i) be an inner product over K and
e H, kek = 1. If (0, 1) , then
(2.227) Re [hx, yi hx, ei he, yi]

1
1

kx + (1 ) yk2 |hx + (1 ) y, ei|2 .
4 (1 )
The constant
1
4
is best possible.
Proof. Firstly, note that the following equality holds true

hx hx, ei e, y hy, ei ei = hx, yi hx, ei he, yi .
Utilising the elementary inequality
1
Re hz, wi kz + wk2 ,
4
we have
z, w H
Re hx hx, ei e, y hy, ei ei
1
=
Re hx hx, ei e, (1 ) y h(1 ) y, ei ei
(1 )

1
1

kx + (1 ) yk2 |hx + (1 ) y, ei|2 ,
4 (1 )
proving the desired inequality (2.227).
Remark 35. For = 12 , we get the simpler inequality:

2
x + y 2 x + y

(2.228)
Re [hx, yi hx, ei he, yi]

, e ,

2
2
96
that has been obtained in [4, p. 46], for which the sharpness of the
inequality was established.
space over K and p 1. Then for any x, y H we have
0 kxk kyk |hx, yi| kxk kyk Re hx, yi

1
(kxk + kyk)2p kx + yk2p p ,
1

2
1

kx yk2p |kxk kyk|2p p .
(2.229)
Proof. Firstly, observe that

2 (kxk kyk Re hx, yi) = (kxk + kyk)2 kx + yk2 .
Denoting D := kxk kyk Re hx, yi , then we have
2D + kx + yk2 = (kxk + kyk)2 .
(2.230)
Taking in (2.230) the power p 1 and using the elementary inequality

(a + b)p ap + bp ; a, b 0,
(2.231)
we have
(kxk + kyk)2p = 2D + kx + yk2
p
2p Dp + kx + yk2p
giving

1
(kxk + kyk)2p kx + yk2p ,
p
2
which is clearly equivalent to the first branch of the third inequality in
(2.229).
With the above notation, we also have
Dp
(2.232)
2D + (kxk kyk)2 = kx yk2 .
Taking the power p 1 in (2.232) and using the inequality (2.231) we

deduce
kx yk2p 2p Dp + |kxk kyk|2p ,
from where we get the last part of (2.229).
97
2.8.3. More Schwarz Related Inequalities. Before we point

out other inequalities related to the Schwarz inequality, we need the
following identity that is interesting in itself [7].
over the real or complex number field K, e H, kek = 1, H and
, K. Then we have the identity:
(2.233) kxk2 |hx, ei|2
= (Re Re hx, ei) (Re hx, ei Re )
+ (Im Im hx, ei) (Im hx, ei Im )

2

1 | |2 .
+
x
e

2
4
Proof. We start with the following known equality (see for instance [5, eq. (2.6)])
(2.234) kxk2 |hx, ei|2
h

i
= Re ( hx, ei) hx, ei Re he x, x ei
holding for x H, e H, kek = 1 and , K.
We also know that (see for instance [14])

2

1
2

(2.235)
Re he x, x ei =
x 2 e 4 | | .
Since
(2.236)
i
h

Re ( hx, ei) hx, ei
= (Re Re hx, ei) (Re hx, ei Re )
+ (Im Im hx, ei) (Im hx, ei Im ) ,
hence, by (2.234) (2.236), we deduce the desired identity (2.233).

The following general result providing a reverse of the Schwarz inequality may be stated [7].
Proposition 32. Let (H; h, i) be an inner product space over K,
e H, kek = 1, x H and , K. Then we have the inequality:

2

+
2
2

(2.237)
(0 ) kxk |hx, ei| x
e
.
2
98
The constant 12 is best possible in (2.237). The case of equality holds

in (2.237) if and only if

+
+
,
Im hx, ei = Im
.
(2.238)
Re hx, ei = Re
2
2
Proof. Utilising the elementary inequality for real numbers
1
( + )2 ,
, R;
4
with equality iff = , we have
1
(2.239)
(Re Re hx, ei) (Re hx, ei Re ) (Re Re )2
4
and
1
(2.240)
(Im Im hx, ei) (Im hx, ei Im ) (Im Im )2
4
with equality if and only if
Re + Re
Im + Im
Re hx, ei =
and
Im hx, ei =
.
2
2
Finally, on making use of (2.239), (2.240) and the identity (2.233), we
deduce the desired result (2.237).
e H, kek = 1, x H and , K. If x H is such that
(2.241)
Re Re hx, ei Re
and
Im Im hx, ei Im ,
(2.242)
1 | |2 .
kxk |hx, ei|
x
e

2
4
2
The constant 41 is best possible in (2.242). The case of equality holds

in (2.242) if and only if
Re hx, ei = Re or Re hx, ei = Re
and
Im hx, ei = Im or Im hx, ei = Im .
Proof. From the hypothesis we obviously have
(Re Re hx, ei) (Re hx, ei Re ) 0
and
(Im Im hx, ei) (Im hx, ei Im ) 0.
99
Utilising the identity (2.233) we deduce the desired result (2.242). The
case of equality is obvious.
Further on, we can state the following reverse of the quadratic
Schwarz inequality [7]:
e H, kek = 1. If , K and x H are such that either
Re he x, x ei 0
(2.243)
or, equivalently,
x
e
| | ,

2 2
(2.244)
then
(2.245)
(0 ) kxk2 |hx, ei|2

(Re Re hx, ei) (Re hx, ei Re )
+ (Im Im hx, ei) (Im hx, ei Im )
1
| |2 .
4
The case of equality holds in (2.245) if it holds either in (2.243) or
(2.244).
The proof is obvious by Lemma 4 and we omit the details.

Remark 36. We remark that the inequality (2.245) may also be
used to get, for instance, the following result
(2.246) kxk2 |hx, ei|2

1
(Re Re hx, ei)2 + (Im Im hx, ei)2 2

1
(Re hx, ei Re )2 + (Im hx, ei Im )2 2 ,
that provides a different bound than
|hx, ei|2 .
1
4
| |2 for the quantity kxk2

space over K and , > 0, K with ||2 . If x, a H are such
that a 6= 0 and
12

2

||
x a
(2.247)
kak ,

100
then we have the following reverses of Schwarzs inequality

Re Re hx, ai + Im Im hx, ai
|| |hx, ai|

kxk kak
(2.248)
and
||2
(0 ) kxk kak |hx, ai|
|hx, ai|2 .
(2.249)
Proof. Taking the square in (2.247), it becomes equivalent to

||2
||2
2
2
kak2 ,
kxk Re hx, ai + 2 kak
2

(2.250)
kxk2 + kak2 2 Re hx, ai
2
= 2 [Re Re hx, ai + Im Im hx, ai] .

On the other hand, since
(2.251)
2 kxk kak kxk2 + kak2 ,
hence by (2.250) and (2.251) we deduce the first inequality in (2.248).
The other inequalities are obvious.
Remark 37. The above inequality (2.248) contains in particular
the reverse (2.191) of the Schwarz inequality. Indeed, if we assume
that = 1, = +
, , K, with = Re (
) > 0, then the
2
2
2
condition || is equivalent to | + | 4 Re (
) which is
2
actually | | 0. With this assumption, (2.247) becomes

x + a 1 | | kak ,

2
2
which implies the reverse of the Schwarz inequality

+ hx, ai
Re
q
kxk kak

2 Re
| + |
q
|hx, ai| ,
2 Re
which is (2.191).
The following particular case of Theorem 38 may be stated [7]:
101
Corollary 20. Let (H; h, i) be an inner product space over K,

[0, 2), 0, 2 . If x, a H are such that a 6= 0 and
(2.252)
kx (cos + i sin ) ak cos kak ,
then we have the reverses of the Schwarz inequality

cos Re hx, ai + sin Im hx, ai
(2.253)
kxk kak
.
sin
In particular, if
kx ak cos kak ,
then
1
kxk kak
Re hx, ai ;
cos
and if
kx iak cos kak ,
then
1
Im hx, ai .
kxk kak
cos
2.8.4. Reverses of the Generalised Triangle Inequality. In
[13], the author obtained the following reverse result for the generalised
triangle inequality

n
n
X

X

(2.254)
kxi k
xi ,

i=1
i=1
provided xi H, i {1, . . . , n} are vectors in a real or complex inner

product (H; h, i) :
Theorem 39 (Dragomir, 2004). Let e, xi H, i {1, . . . , n} with
kek = 1. If ki 0, i {1, . . . , n} are such that
(2.255) (0 ) kxi k Re he, xi i ki
for each

n
n
n
X
X
X

(2.256)
(0 )
kxi k
xi
ki .

i=1
i=1
i=1
The equality holds in (2.256) if and only if

n
n
X
X
(2.257)
kxi k
ki
i=1
i=1
and
(2.258)
n
X
i=1
xi =
n
X
i=1
kxi k
n
X
i=1
!
ki e.
i {1, . . . , n} ,
102
By utilising some of the results obtained in Section 2.8.2, we point

out several reverses of the generalised triangle inequality (2.254) that
are corollaries of the above Theorem 39 [7].
Corollary 21. Let e, xi H\ {0} , i {1, . . . , n} with kek = 1.
If

xi
ri

e

kxi k
(2.259)
for each
i {1, . . . , n} ,
then
(2.260)
(0 )
n
X

n
X

kxi k
xi

i=1
n
X
1
2
i=1
ri2 kxi k
i=1

2 n
P
max ri
kxi k ;
1in
i=1
p1 n
1q
1 P
n
P
q
2p

, p > 1,
ri
kxi k
2
i=1
i=1
n
P
max kxi k ri2 .

1in
1
p
1
q
= 1;
i=1
Proof. The first part follows from Proposition 25 on choosing x =

xi , y = e and applying Theorem 39. The last part is obvious by Holders
inequality.
Remark 38. One would obtain the same reverse inequality (2.260)
if one were to use Theorem 26. In this case, the assumption (2.259)
should be replaced by
(2.261)
kkxi k xi ek ri kxi k
for each
i {1, . . . , n} .
On utilising the inequalities (2.198) and (2.209) one may state the
following corollary of Theorem 39 [7].
Corollary 22. Let e, xi H\ {0} , i {1, . . . , n} with kek = 1.
Then we have the inequality

n
n
X

X

(2.262)
(0 )
kxi k
xi min {A, B} ,

i=1
i=1
103
where
A := 2
n
X
i=1

kxi k
kxi ek
kxi k + 1
2
,
and
n
1 X (kxi k + 1)2 kxi ek2

B :=
.
2 i=1
kxi k
(0, 1) := {z H| kzk 1} ,
For vectors located outside the closed unit ball B
we may state the following result [7].
(0, 1) , i {1, . . . , n} and
Corollary 23. Assume that xi
/ B
e H, kek = 1. Then we have the inequality:

n
n
X

X

(0 )
(2.263)
kxi k
xi

i=1
i=1
1 P
n
2
p
kx kp1 kxi ek2 , if p 1
2 i=1 i
1P
kxi k1p kxi ek2 ,

if p < 1.
2 i=1
The proof follows by Proposition 27 and Theorem 39.
For complex spaces one may state the following result as well [7].
Corollary 24. Let (H; h, i) be a complex inner product space
and i C with Re i , Im i > 0, i {1, . . . , n} . If xi , e H, i
{1, . . . , n} with kek = 1 and

Im
i
xi
di ,
(2.264)
e
i {1, . . . , n} ,

Re i
then
(2.265)
(0 )
n
X
i=1

n
n
X
1X
Re i 2

kxi k
xi
d .

2
Im i i
i=1
i=1
The proof follows by Theorems 30 and 39 and the details are omitted.
Finally, by the use of Theorem 37, we can state [7]:
104
Corollary 25. If xi , e H, i {1, . . . , n} with kek = 1 and

p 1, then we have the inequalities:

n
n
X
X

(2.266)
(0 )
kxi k
xi

i=1
i=1
P
n
1
2p
2p p
(kx
k
+
1)
kx
+
ek
,
i
i
1 i=1

2
n
1
P
kxi ek2p |kxi k 1|2p p .

i=1
Bibliography
[1] C. BLATTER, Zur Riemannschen Geometrie im Grossen auf dem Mobiusband.
(German) Compositio Math. 15 (1961), 88107.
[2] N.G. de BRUIJN, Problem 12, Wisk. Opgaven, 21 (1960), 12-14.
[3] M.L. BUZANO, Generalizzazione della diseguaglianza di Cauchy-Schwarz.
(Italian), Rend. Sem. Mat. Univ. e Politech. Torino, 31 (1971/73), 405409
(1974).
[4] S.S. DRAGOMIR, Advances in Inequalities of the Schwarz, Gr
uss and Bessel
Type in Inner Product Spaces, RGMIA Monographs, Victoria University, 2004.
[ONLINE: http://rgmia.vu.edu.au/monographs/].
[5] S.S. DRAGOMIR, A generalisation of Gr
uss inequality in inner product spaces
and applications, J. Mathematical Analysis and Applications, 237 (1999), 7482.
[6] S.S. DRAGOMIR, A generalisation of Kurepas inequality, RGMIA Res. Rep.
Coll., 7(E) (2004), Art. 23. [ONLINE http://rgmia.vu.edu.au/v7(E).html]
[7] S.S. DRAGOMIR, A potpourri of Schwarz related inequalities in inner product
spaces, ArXiv:math. MG/0501129v1 [ONLINE ]
[8] S.S. DRAGOMIR, A refinement of Cauchy-Schwarzs inequality, Gazeta Mat.
Metod. (Bucharest, Romania), 8 (1987), 94-95.
[9] S.S. DRAGOMIR, Discrete Inequalities of the Cauchy-Bunyakovsky-Schwarz
Type, Nova Science Publishers, NY, 2004.
[10] S.S. DRAGOMIR, Generalizations of Precupanus inequality for orthornormal
families of vectors in inner product spaces, RGMIA Res. Rep. Coll., 7(E)
(2004), Art. 26. [ONLINE http://rgmia.vu.edu.au/v7(E).html]
[11] S.S. DRAGOMIR, Inequalities for orthornormal families of vectors in
inner product spaces related to Buzanos, Richards and Kurepas
results, RGMIA Res. Rep. Coll., 7(E) (2004), Art. 25. [ONLINE
http://rgmia.vu.edu.au/v7(E).html]
[12] S.S. DRAGOMIR, Refinements of the Schwarz and Heisenberg inequalities in
Hilbert spaces, J. Inequal. Pure & Appl. Math., 5(3) (2004), Art. 60. [ONLINE:
http://jipam.vu.edu.au/article.php?sid=446]
[13] S.S. DRAGOMIR, Reverses of the triangle inequality in inner product spaces, RGMIA Res. Rep. Coll., 7(E) (2004), Article 7. [ONLINE:
http://rgmia.vu.edu.au/v7(E).html].
[14] S.S. DRAGOMIR, Some Gr
uss type inequalities in inner product spaces,
J. Inequal. Pure & Appl. Math., 4(2) (2003), Article 42. [Online:
http://jipam.vu.edu.au/article.php?sid=280].
[15] S.S. DRAGOMIR, Some refinements of Schwartz inequality, Simpozionul de
Matematici si Aplicatii, Timisoara, Romania, 1-2 Noiembrie 1985, 1316.
105
106
BIBLIOGRAPHY
[16] S.S. DRAGOMIR, Refinements of Buzanos and Kurepas inequalities in inner product spaces, RGMIA Res. Rep. Coll., 7(E) (2004), Art. 24. [ONLINE
http://rgmia.vu.edu.au/v7(E).html]
[17] S.S. DRAGOMIR and B. MOND, On the superadditivity and monotonicity of
Schwarzs inequality in inner product spaces, Contributios, Macedonian Acad.
Sci. Arts., 15(2) (1994), 5-22.

Some inequalities in prehilbertian spaces,
Studia Univ., Babes-Bolyai, Mathematica, 32(1)(1987), 71-78 MR 89h: 46034.

Some inequalities in prehilbertian spaces,
Conferinta National
a de Geometrie si Topologie, Targoviste, Romania, 12-14
Aprilie, 1986, 73-76.
[20] C.F. DUNKL and K.S. WILLIAMS, A simple norm inequality, The Amer.
Math. Monthly, 71(1) (1964), 43-54.
[21] M. FUJII and F. KUBO, Buzanos inequality and bounds for roots of algebraic
equations, Proc. Amer. Math. Soc., 117(2) (1993), 359-361.
[22] A.A. GOLDSTEIN, J.V. RYFF and L.E. CLARKE, Problem 5473, The Amer.
Math. Monthly, 75(3) (1968), 309.
[23] G.H. HARDY, J.E. LITTLEWOOD and G. POLYA, Inequalities, Cambridge
University Press, Cambridge, United Kingdom, 1952.
[24] G.N. HILE, Entire solution of linear elliptic equations with Laplacian principal
part, Pacific J. Math., 62 (1976), 127-148.
[25] S. KUREPA, On the Buniakowsky-Cauchy-Schwarz inequality, Glasnick Mathematicki, 1(21)(2) (1966), 147-158.
Analytic Inequalities, Springer Verlag, 1970.
[26] D.S. MITRINOVIC,
[27] M.H. MOORE, An inner product inequality, SIAM J. Math. Anal., 4(1973),
No. 3, 514-518.
On some classical inequalities in unitary spaces, Mat. Bilten,

[28] J.E. PECARI
C,
(Macedonia) 16 (1992), 63-72.
[29] T. PRECUPANU, On a generalisation of Cauchy-Buniakowski-Schwarz inequality, Anal. St. Univ. Al. I. Cuza Iasi, 22(2) (1976), 173-175.
[30] U. RICHARD, Sur des inegalites du type Wirtinger et leurs application aux
equations differentielles ordinaires, Collquium of Analysis held in Rio de
Janeiro, August, 1972, pp. 233-244.
CHAPTER 3
Reverses for the Triangle Inequality

3.1. Introduction
The following reverse of the generalised triangle inequality

n
n
X

X

cos
|zk |
zk ,

k=1
k=1
provided the complex numbers zk , k {1, . . . , n} satisfy the assumption

a arg (zk ) a + , for any k {1, . . . , n} ,

where a R and 0, 2 was first discovered by M. Petrovich in
1917, [11] (see [10, p. 492]) and subsequently was rediscovered by
other authors, including J. Karamata [6, p. 300 301], H.S. Wilf [12],
and in an equivalent form by M. Marden [8].
In 1966, J.B. Diaz and F.T. Metcalf [1] proved the following reverse
of the triangle inequality:
Theorem 40 (Diaz-Metcalf, 1966). Let a be a unit vector in the
inner product space (H; h, i) over the real or complex number field K.
Suppose that the vectors xi H\ {0} , i {1, . . . , n} satisfy
(3.1)
0r
Re hxi , ai
,
kxi k
i {1, . . . , n} .
Then
(3.2)
n
X
i=1

n
X

kxi k
xi ,

i=1
where equality holds if and only if

(3.3)
n
X
i=1
xi = r
n
X
!
kxi k a.
i=1
A generalisation of this result for orthonormal families is incorporated in the following result [1].
107
108
3. REVERSES FOR THE TRIANGLE INEQUALITY
Theorem 41 (Diaz-Metcalf, 1966). Let a1 , . . . , an be orthonormal

vectors in H. Suppose the vectors x1 , . . . , xn H\ {0} satisfy
(3.4)
0 rk
Re hxi , ak i
,
kxi k
i {1, . . . , n} , k {1, . . . , m} .
Then
(3.5)
m
X
! 21
rk2
n
X
i=1
k=1
n

X

kxi k
xi ,

i=1

(3.6)
n
X
xi =
i=1
n
X
!
kxi k
i=1
m
X
r k ak .
k=1
Similar results valid for semi-inner products may be found in [7]

and [9].
For other classical inequalities related to the triangle inequality, see
Chapter XVII of the book [10] and the references therein.
The aim of the present chapter is to provide various recent reverses
for the generalised triangle inequality in both its simple form that are
closely related to the Diaz-Metcalf results mentioned above, or in the
equivalent quadratic form, i.e., upper bounds for
n
X
i=1
!2
kxi k

2
n
X

xi

i=1
and
P
2
k ni=1 xi k
P
2.
( ni=1 kxi k)
Applications for vector valued integral inequalities and for complex
numbers are given as well.
3.2. Some Inequalities of Diaz-Metcalf Type
3.2.1. The Case of One Vector. The following result with a
natural geometrical meaning holds [3]:
Theorem 42 (Dragomir, 2004). Let a be a unit vector in the inner
product space (H; h, i) and (0, 1) . If xi H, i {1, . . . , n} are
such that
(3.7)
kxi ak for each i {1, . . . , n} ,
3.2. SOME INEQUALITIES OF DIAZ-METCALF TYPE
109

(3.8)

n
n
X

X
p

2
1
kxi k
xi ,

i=1
i=1

(3.9)
n
X
xi =
1 2
i=1
n
X
!
kxi k a.
i=1
Proof. From (3.7) we have

kxi k2 2 Re hxi , ai + 1 2 ,
giving
(3.10)
kxi k2 + 1 2 2 Re hxi , ai ,
for each i {1, .p

. . , n} .
Dividing by 1 2 > 0, we deduce
(3.11)
p
kx k2
2 Re hxi , ai
p i
+ 1 2 p
,
1 2
1 2
for each i {1, . . . , n} .

On the other hand, by the elementary inequality
p
(3.12)
+ q 2 pq, p, q 0, > 0
we have
p
kxi k2
(3.13)
2 kxi k p
+ 1 2
1 2
and thus, by (3.11) and (3.13), we deduce
Re hxi , ai p
1 2 ,
kxi k
for each i {1, . . . , n} . Applying Theorem 40 for r =
deduce the desired inequality (3.8).
1 2 , we
The following results may be stated as well [3].

Theorem 43 (Dragomir, 2004). Let a be a unit vector in the inner
product space (H; h, i) and M m > 0. If xi H, i {1, . . . , n} are
such that either
(3.14)
Re hM a xi , xi mai 0
110
or, equivalently,

1
M
+
m
xi
(M m)
a

2
2
(3.15)
holds for each i {1, . . . , n} , then we have the inequality

n
n
X

2 mM X

(3.16)
kxi k
xi ,

m + M i=1
i=1
or, equivalently,
(3.17)
(0 )
n
X
i=1
2
n

X
M
kxi k
xi

2 mM
i=1
n

X

xi .

i=1
The equality holds in (3.16) (or in (3.17)) if and only if

!
n
n
X
2 mM X
(3.18)
xi =
kxi k a.
m
+
M
i=1
i=1
Proof. Firstly, we remark that if x, z, Z H, then the following
statements are equivalent:
(i) Re
x,
hZ Z+z
x 1 zi 0;

(ii) x 2 2 kZ zk .
Using this fact, one may simply realize that (3.14) and (3.15) are
equivalent.
Now, from (3.14), we get
kxi k2 + mM (M + m) Re hxi , ai ,
for any i {1, . . . , n} . Dividing this inequality by mM > 0, we

deduce the following inequality that will be used in the sequel
M +m
kxi k2
Re hxi , ai ,
+ mM
(3.19)
mM
mM
for each i {1, . . . , n} .
Using the inequality (3.12) from Theorem 42, we also have
kxi k2
+ mM ,
2 kxi k
mM
for each i {1, . . . , n} .
Utilizing (3.19) and (3.20), we may conclude with the following
inequality
M +m
Re hxi , ai ,
kxi k
mM
(3.20)
3.2. SOME INEQUALITIES OF DIAZ-METCALF TYPE
which is equivalent to
(3.21)
111
2 mM
Re hxi , ai
m+M
kxi k
for any i {1, . . . , n} .

Finally, on applying the Diaz-Metcalf result in Theorem 40 for r =
2 mM
, we deduce the desired conclusion.
m+M
The equivalence between (3.16) and (3.17) follows by simple calculation and we omit the details.
3.2.2. The Case of m Vectors. In a similar manner to the one
used in the proof of Theorem 42 and by the use of the Diaz-Metcalf
inequality incorporated in Theorem 41, we can also prove the following
result [3] :
Proposition 35. Let a1 , . . . , an be orthonormal vectors in H. Suppose the vectors x1 , . . . , xn H\ {0} satisfy
(3.22)
kxi ak k k for each i {1, . . . , n} , k {1, . . . , m} ,
where k (0, 1) , k {1, . . . , m} . Then we have the following reverse

of the triangle inequality
n

! 12 n
m
X
X
X

kxi k
xi .
(3.23)
m
2k

i=1
k=1
i=1

! m
n
n
X
X
X
1
(3.24)
xi =
kxi k
1 2k 2 ak .
i=1
i=1
k=1
Finally, by the use of Theorem 41 and a similar technique to that

employed in the proof of Theorem 43, we may state the following result
[3]:
Proposition 36. Let a1 , . . . , an be orthonormal vectors in H. Suppose the vectors x1 , . . . , xn H\ {0} satisfy
(3.25)
Re hMk ak xi , xi k ak i 0,
or, equivalently,
(3.26)

1
M
+
k
k
xi
(Mk k ) ,
a
k

2
2
for any i {1, . . . , n} and k {1, . . . , m} , where Mk k > 0 for

each k {1, . . . , m} .
112

(3.27)
m
X
k=1
k M k
(k + Mk )2
! 12
n
X
i=1
n

X

kxi k
xi .

i=1
The equality holds in (3.27) iff

(3.28)
n
X
n
X
xi = 2
i=1
!
kxi k
i=1
m
X
k=1
p
k M k
ak .
k + M k
3.3. Additive Reverses for the Triangle Inequality

3.3.1. The Case of One Vector. In this section we establish
some additive reverses of the generalised triangle inequality in real or
complex inner product spaces.
The following result holds [3]:
space over the real or complex number field K and e, xi H, i
{1, . . . , n} with kek = 1. If ki 0, i {1, . . . , n} , are such that
(3.29)
kxi k Re he, xi i ki for each i {1, . . . , n} ,

(3.30)
(0 )
n
X
i=1

n
n
X
X

kxi k
xi
ki .

i=1
i=1

n
X
(3.31)
kxi k
i=1
n
X
ki
i=1
and
(3.32)
n
X
i=1
xi =
n
X
kxi k
i=1
n
X
!
ki e.
i=1
Proof. If we sum in (3.29) over i from 1 to n, then we get

* n
+
n
n
X
X
X
(3.33)
kxi k Re e,
xi +
ki .
i=1
i=1
i=1
3.3. ADDITIVE REVERSES FOR THE TRIANGLE INEQUALITY
113
P
By Schwarzs inequality for e and ni=1 xi , we have
* n
+ * n
+

X
X

(3.34) Re e,
xi Re e,
xi

i=1
i=1
*
n
n

+
n

X X

X

xi =
xi .
e,
xi kek

i=1
i=1
i=1
Making use of (3.33) and (3.34), we deduce the desired inequality

(3.29).
If (3.31) and (3.32) hold, then
n
n

n
n
n
X X

X
X
X

xi =
kxi k
ki kek =
kxi k
ki ,

i=1
i=1
i=1
i=1
i=1
and the equality in the second part of (3.30) holds true.

Conversely, if the equality holds in (3.30), then, obviously (3.31) is
valid and we need only to prove (3.32).
Now, if the equality holds in (3.30) then it must hold in (3.29) for
each i {1, . . . , n} and also must hold in any of the inequalities in
(3.34).
It is well known that in Schwarzs inequality |hu, vi| kuk kvk
(u, v H) the case of equality holds iff there exists a K such that
u = v. We note that in the weaker inequality Re hu, vi kuk kvk the
case of equality holds iff 0 and u = v.
Consequently, the equality holds in all inequalities (3.34) simultaneously iff there exists a 0 with
(3.35)
e =
n
X
xi .
i=1
If we sum the equalities in (3.29) over i from 1 to n, then we deduce

* n
+
n
n
X
X
X
(3.36)
kxi k Re e,
xi =
ki .
i=1
Replacing
Pn
i=1
i=1
i=1
kxi k from (3.35) into (3.36), we deduce

n
X
kxi k kek2 =
n
X
i=1
Pn
from where we get = i=1 kxi k

(3.32) and the theorem is proved.
ki ,
i=1
Pn
i=1
ki . Using (3.35), we deduce
114
3.3.2. The Case of m Vectors. If we turn our attention to the

case of orthogonal families, then we may state the following result as
well [3].
space over the real or complex number field K, {ek }k{1,...,m} a family
of orthonormal vectors in H, xi H, Mi,k 0 for i {1, . . . , n} and
k {1, . . . , m} such that
kxi k Re hek , xi i Mik
(3.37)
for each i {1, . . . , n} , k {1, . . . , m} . Then we have the inequality
(3.38)
n
X
i=1

n
n
m

1
X 1 X X
xi +
Mik .
kxi k
m i=1 m i=1 k=1
The equality holds true in (3.38) if and only if

n
X
(3.39)
kxi k
i=1
1 XX
Mik
m i=1 k=1
and
(3.40)
n
X
xi =
i=1
n
X
i=1
1 XX
kxi k
Mik
m i=1 k=1
m
X
ek .
k=1
Proof. If we sum over i from 1 to n in (3.37), then we obtain

n
X
i=1
*
kxi k Re e,
n
X
i=1
+
xi
n
X
Mik ,
i=1
for each k {1, . . . , m} . Summing these inequalities over k from 1 to

m, we deduce
(3.41)
n
X
i=1
* m
+
n
n
m
X X
1
1 XX
kxi k
Re
ek ,
xi +
Mik .
m
m
i=1
i=1 k=1
k=1
3.3. ADDITIVE REVERSES FOR THE TRIANGLE INEQUALITY
115
Pn
P
By Schwarzs inequality for m
k=1 ek and
i=1 xi we have
* m
+ * m
+
n
n

X X
X
X

Re
ek ,
xi Re
(3.42)
ek ,
xi

i=1
k=1
* m k=1 n i=1+
X X

ek ,
xi

k=1 i=1
m
n
X
X

ek
xi

i=1
k=1

n

X
xi ,
= m

i=1
since, obviously,

v

2 v
u m
m
m
X
u
X

u
uX

t

ek = t
ek =
kek k2 = m.

k=1
k=1
k=1

(3.38).

n
n
m
m
n
X

1
1 XX
X X

xi =
kxi k
Mik
ek

m i=1 k=1
m i=1 i=1
k=1
!
n
n
m
m X
1 XX
=
Mik
kxi k
m i=1 k=1
m i=1
=
n
X
i=1
1 XX
kxi k
Mik ,
m i=1 k=1
and the equality in (3.38) holds true.

valid.
Now if the equality holds in (3.38), then it must hold in (3.37) for
each i {1, . . . , n} and k {1, . . . , m} and also must hold in any of
the inequalities in (3.42).
It is well known that in Schwarzs inequality Re hu, vi kuk kvk ,
the equality occurs iff u = v with 0, consequently, the equality
holds in all inequalities (3.42) simultaneously iff there exists a 0
116
with
(3.43)
m
X
ek =
n
X
xi .
i=1
k=1
If we sum the equality in (3.37) over i from 1 to n and k from 1 to m,

then we deduce
+
* m
n
n X
m
n
X X
X
X
Mik .
kxi k Re
ek ,
xi =
(3.44)
m
i=1
Replacing
Pn
i=1
k=1
i=1 k=1
xi from (3.43) into (3.44), we deduce
i=1
n
X
kxi k
i=1
m
X
kek k =
n X
m
X
Mik
i=1 k=1
k=1
giving
=
n
X
i=1
1 XX
kxi k
Mik .
m i=1 k=1
Using (3.43), we deduce (3.40) and the theorem is proved.

3.4. Further Additive Reverses
3.4.1. The Case of Small Balls. In this section we point out
different additive reverses of the generalised triangle inequality under
simpler conditions for the vectors involved.
Theorem 46 (Dragomir, 2004). Let (H; h, i) be an inner product space over the real or complex number field K and e, xi H,
i {1, . . . , n} with kek = 1. If (0, 1) and xi , i {1, . . . , n} are
such that
(3.45)
kxi ek for each i {1, . . . , n} ,

(3.46)
(0 )
n
X
i=1
n

X

kxi k
xi

i=1
2
1
1+
Re
* n
X
+
xi , e
i=1

n

X

p

xi .
p

1 2 1 + 1 2
i=1
3.4. FURTHER ADDITIVE REVERSES
117

n
X
(3.47)
i=1
kxi k p
1+
Re
* n
X
+
xi , e
i=1
and
(3.48)
n
X
xi
i=1
n
X
i=1
kxi k p
1+
Re
* n
X
+
xi , e e.
i=1
Proof. We know, from the proof of Theorem 44, that, if (3.45) is

fulfilled, then we have the inequality
1
kxi k p
Re hxi , ei
1 2
for each i {1, . . . , n} , implying
(3.49)
kxi k Re hxi , ei
1
p
1 2
1 Re hxi , ei
2
=p
1 2 1 +
1 2
for each i {1, . . . , n} .

Now, making use of Theorem 42, for
2

Re hxi , ei ,
ki := p
p
1 2 1 + 1 2
Re hxi , ei
i {1, . . . , n} ,
we easily deduce the conclusion of the theorem.

We omit the details.
We may state the following result as well [3]:
space and e H, M m > 0. If xi H, i {1, . . . , n} are such that
either
(3.50)
Re hM e xi , xi mei 0,
or, equivalently,
(3.51)

xi M + m e 1 (M m)

2
2
118
holds for each i {1, . . . , n} , then we have the inequality

2

* n
+
n
n

M m
X
X
X
kxi k
xi
Re
(3.52) (0 )
xi , e

2 mM
i=1
i=1
i=1

2

n

M m X
xi .

2 mM
i=1

2
* n
+
n
M m
X
X
(3.53)
kxi k
Re
xi , e
2
mM
i=1
i=1
and
(3.54)
n
X
i=1
xi =
n
X
i=1

kxi k
2
* n
+
M m
X
Re
xi , e e.
2 mM
i=1
Proof. We know, from the proof of Theorem 43, that if (3.50) is

fulfilled, then we have the inequality
M +m
Re hxi , ei
kxi k
2 mM
for each i {1, . . . , n} . This is equivalent to

2
M m
kxi k Re hxi , ei
Re hxi , ei
2 mM
for each i {1, . . . , n} .
Now, making use of Theorem 44, we deduce the conclusion of the
theorem. We omit the details.
Remark 39. If one uses Theorem 45 instead of Theorem 44 above,
then one can state the corresponding generalisation for families of orthonormal vectors of the inequalities (3.46) and (3.52) respectively. We
do not provide them here.
3.4.2. The Case of Arbitrary Balls. Now, on utilising a slightly
different approach, we may point out the following result [3]:
space over K and e, xi H, i {1, . . . , n} with kek = 1. If ri > 0,
3.4. FURTHER ADDITIVE REVERSES
119
i {1, . . . , n} are such that

(3.55)
kxi ek ri for each i {1, . . . , n} ,

(3.56)
n
X
i=1

n
n
X
1X

kxi k
ri2 .
xi

2
i=1
i=1

n
n
X
1X 2
(3.57)
kxi k
ri
2
i=1
i=1
and
(3.58)
n
X
i=1
xi =
n
X
i=1
1X 2
kxi k
r
2 i=1 i
!
e.
Proof. The condition (3.55) is clearly equivalent to

(3.59)
kxi k2 + 1 Re hxi , ei + ri2
for each i {1, . . . , n} .

Using the elementary inequality
(3.60)
2 kxi k kxi k2 + 1,
for each i {1, . . . , n} , then, by (3.59) and (3.60), we deduce

2 kxi k 2 Re hxi , ei + ri2 ,
giving
(3.61)
1
kxi k Re hxi , ei ri2
2
for each i {1, . . . , n} .

Now, utilising Theorem 44 for ki = 21 ri2 , i {1, . . . , n} , we deduce
the desired result. We omit the details.
Finally, we may state and prove the following result as well [3].
space over K and e, xi H, i {1, . . . , n} with kek = 1. If Mi mi >
0, i {1, . . . , n} , are such that

1
M
+
m
i
i
xi
(Mi mi ) ,
(3.62)
e

2
2
or, equivalently,
(3.63)
Re hMi e x, x mi ei 0
120
for each i {1, . . . , n} , then we have the inequality

n
n
n
X
1X
X
(Mi mi )2

(3.64)
(0 )
kxi k
xi
.

4
Mi + m i
i=1
i=1
i=1

n
X
(3.65)
i=1
1 X (Mi mi )2
kxi k
4 i=1 Mi + mi
and
n
X
(3.66)
xi =
i=1
n
X
i=1
1 X (Mi mi )2
kxi k
4 i=1 Mi + mi
!
e.
Proof. The condition (3.62) is equivalent to:

2
kxi k +
Mi + m i
2
2

Mi + m i
1
2 Re xi ,
e + (Mi mi )2
2
4
and since

2
Mi + m i
2
kxi k kxi k +
Mi + m i
2
2
,
then we get

Mi + m i
Mi + m i
1
kxi k 2
Re hxi , ei + (Mi mi )2 ,
2
2
2
4
or, equivalently,
1 (Mi mi )2
kxi k Re hxi , ei
4
Mi + m i
for each i {1, . . . , n} .
2
i mi )
Now, making use of Theorem 44 for ki := 41 (M
, i {1, . . . , n} ,
Mi +mi
we deduce the desired result.
Remark 40. If one uses Theorem 45 instead of Theorem 44 above,
then one can state the corresponding generalisation for families of orthonormal vectors of the inequalities in (3.56) and (3.64) respectively.
3.5. REVERSES OF SCHWARZ INEQUALITY
121
3.5. Reverses of Schwarz Inequality

In this section we outline a procedure showing how some of the
above results for triangle inequality may be employed to obtain reverses
for the celebrated Schwarz inequality.
For a H, kak = 1 and r (0, 1) define the closed ball
D (a, r) := {x H, kx ak r} .
The following reverse of the Schwarz inequality holds [3]:
Proposition 37. If x, y D (a, r) with a H, kak = 1 and
r (0, 1) , then we have the inequality
(3.67)
(0 )
kxk kyk Re hx, yi

1
r2 .
2
2
(kxk + kyk)
The constant 12 in (3.67) is best possible in the sense that it cannot be

replaced by a smaller quantity.
Proof. Using Theorem 42 for x1 = x, x2 = y, = r, we have
(3.68)
1 r2 (kxk + kyk) kx + yk .
Taking the square in (3.68) we deduce

1 r2 kxk2 + 2 kxk kyk + kyk2 kxk2 + 2 Re hx, yi + kyk2
Now, assume that (3.67) holds with a constant C > 0 instead of
1
, i.e.,
2
(3.69)
kxk kyk Re hx, yi

Cr2
(kxk + kyk)2
provided x, y D (a, r) with a H, kak = 1 and r (0, 1) .

Let e H with kek = 1 and e a. Define x = a + re, y = a re.
Then
kxk = 1 + r2 = kyk , Re hx, yi = 1 r2

and thus, from (3.69), we have
1 + r2 (1 r2 )
2
2 Cr
2
2 1+r
giving

1
1 + r2 C
2
for any r (0, 1) . If in this inequality we let r 0+, then we get
C 12 and the proposition is proved.
122
In a similar way, by the use of Theorem 43, we may prove the

following reverse of the Schwarz inequality as well [3]:
Proposition 38. If a H, kak = 1, M m > 0 and x, y H
are so that either
Re hM a x, x mai , Re hM a y, y mai 0
or, equivalently,

1
m
+
M
m
+
M
x
a
,
y
a

2 (M m)
2
2
hold, then
kxk kyk Re hx, yi
1
(0 )
2
2
(kxk + kyk)
The constant
1
2
M m
M +m
2
.
cannot be replaced by a smaller quantity.
Remark 41. On utilising Theorem 35 and Theorem 36, we may

deduce some similar reverses of Schwarz inequality provided x, y
m
k=1 D (ak , k ) , assumed not to be empty, where a1 , ..., an are orthonormal vectors in H and k (0, 1) for k {1, ..., m} . We omit the details.
Remark 42. For various different reverses of Schwarz inequality
in inner product spaces, see the recent survey [2].
3.6.1. The General Case. The following lemma holds [4]:
over the real or complex number field K, xi H, i {1, . . . , n} and
kij > 0 for 1 i < j n such that
(3.70)
0 kxi k kxj k Re hxi , xj i kij
for 1 i < j n. Then we have the following quadratic reverse of the

triangle inequality

!2 n
n
X 2
X
X

(3.71)
kxi k
xi + 2
kij .

i=1
i=1
1i<jn
The case of equality holds in (3.71) if and only if it holds in (3.70) for
each i, j with 1 i < j n.
3.6. QUADRATIC REVERSES OF THE TRIANGLE INEQUALITY
123
Proof. We observe that the following identity holds:

!2 n
n
X 2
X

(3.72)
kxi k
xi

i=1
i=1
+
*
n
n
n
X
X
X
=
kxi k kxj k
xi ,
xj
i,j=1
n
X
i,j=1
n
X
kxi k kxj k
i=1
n
X
j=1
Re hxi , xj i
i,j=1
[kxi k kxj k Re hxi , xj i]
i,j=1
1i<jn
1j<in
=2
[kxi k kxj k Re hxi , xj i] .
1i<jn
Using the condition (3.70), we deduce that

X
1i<jn
kij ,
1i<jn
and by (3.72), we get the desired inequality (3.71).

The case of equality is obvious by the identity (3.72) and we omit
the details.
Remark 43. From (3.71) one may deduce the coarser inequality
that might be useful in some applications:

n
n
X

X

0
kxi k
xi

i=1
i=1
1
!
!2
X
X p
kij .
kij
2
2
1i<jn
1i<jn
Remark 44. If the condition (3.70) is replaced with the following

refinement of Schwarzs inequality:
(3.73)
(0 ) ij kxi k kxj k Re hxi , xj i for 1 i < j n,
124
then the following refinement of the quadratic generalised triangle inequality is valid:

2
!2 n
n
n
X 2
X

X
X

(3.74)
kxi k
xi + 2
ij
xi .

i=1
i=1
1i<jn
i=1
The equality holds in the first part of (3.74) iff the case of equality holds
in (3.73) for each 1 i < j n.
Proposition 39. Let (H; h, i) be as above, xi H, i {1, . . . , n}
and r > 0 such that
kxi xj k r
(3.75)
for 1 i < j n. Then
(3.76)
n
X
!2
kxi k
i=1

2
n
X

n (n 1) 2

r .

xi +

2
i=1

1
kxi k kxj k Re hxi , xj i = r2
2
for each i, j with 1 i < j n.
(3.77)
Proof. The inequality (3.75) is obviously equivalent to

kxi k2 + kxj k2 2 Re hxi , xj i + r2
for 1 i < j n. Since
2 kxi k kxj k kxi k2 + kxj k2 , 1 i < j n;
hence
1
kxi k kxj k Re hxi , xj i r2
2
for any i, j with 1 i < j n.
Applying Lemma 5 for kij := 12 r2 and taking into account that
(3.78)
X
1i<jn
kij =
n (n 1) 2
r ,
4
we deduce the desired inequality (3.76). The case of equality is also

obvious by the above lemma and we omit the details.
125
3.6.2. Inequalities in Terms of the Forward Difference. In

the same spirit, and if some information about the forward difference
xk := xk+1 xk (1 k n 1) are available, then the following
simple quadratic reverse of the generalised triangle inequality may be
stated [4].
Corollary 26. Let (H; h, i) be an inner product space and xi
H, i {1, . . . , n} . Then we have the inequality

!2 n
n
n1
X 2 n (n 1) X
X

(3.79)
kxi k
xi +
kxk k .

2
i=1
i=1
k=1
in general by a smaller quantity.
Proof. Let 1 i < j n. Then, obviously,
j1
j1
n1
X
X
X

kxj xi k =
xk
kxk k
kxk k .

k=i
k=i
k=1
Pn1
Applying Proposition 39 for r := k=1 kxk k , we deduce the desired

result (3.79).
To prove the sharpness of the constant 12 , assume that the inequality
(3.79) holds with a constant c > 0, i.e.,

!2 n
n
n1
X 2
X
X

(3.80)
kxi k
xi + cn (n 1)
kxk k

i=1
i=1
k=1
for n 2, xi H, i {1, . . . , n} .
If we choose in (3.80), n = 2, x1 = 12 e, x2 = 12 e, e H, kek = 1,
then we get 1 2c, giving c 21 .
The following result providing a reverse of the quadratic generalised
triangle inequality in terms of the sup-norm of the forward differences
also holds [4].
Proposition 40. Let (H; h, i) be an inner product space and xi
H, i {1, . . . , n} . Then we have the inequality

!2 n
n
X 2 n2 (n2 1)
X

(3.81)
kxi k
xi +
max kxk k2 .
1kn1

12
i=1
The constant
1
12
i=1
126
Proof. As above, we have that

kxj xi k
j1
X
kxk k (j i) max kxk k ,

1kn1
k=i
for 1 i < j n.
Squaring the above inequality, we get
kxj k2 + kxi k2 2 Re hxi , xj i + (j i)2 max kxk k2
1kn1
for any i, j with 1 i < j n, and since

2 kxi k kxj k kxj k2 + kxi k2 ,
hence
0 kxi k kxj k Re hxi , xj i
(3.82)

Applying Lemma 5 for kij :=
1
2
1
(j i)2 max kxk k2
1kn1
2
(j i)2 max kxk k2 , we can

1kn1
state that
n
X
!2
kxi k
i=1
n

X 2

xi +

i=1
X
1i<jn
(j i)2 max kxk k2 .

1kn1
However,
n
n
X
1X
2
(j i) =
(j i) = n
k2
2
i,j=1
1i<jn
k=1
n
X
!2
k
k=1
n2 (n2 1)
=
12
giving the desired inequality.
To prove the sharpness of the constant, assume that (3.81) holds
with a constant D > 0, i.e.,

!2 n
n
X 2
X

(3.83)
kxi k
xi + Dn2 n2 1 max kxk k2
1kn1

i=1
i=1
for n 2, xi H, i {1, . . . , n} .
If in (3.83) we choose n = 2, x1 = 12 e, x2 = 12 e, e H, kek = 1,
1
then we get 1 12D giving D 12
.
127
Proposition 41. Let (H; h, i) be an inner product space and xi

H, i {1, . . . , n} . Then we have the inequality:

!2 n
! p2
n
n1
X 2
X
X
X
2

,
(3.84)
kxi k
xi +
(j i) q
kxk kp

i=1
i=1
1
p
1i<jn
k=1
1
q
where p > 1, + = 1.
The constant E = 1 in front of the double sum cannot generally be
replaced by a smaller constant.
Proof. Using Holders inequality, we have
kxj xi k
j1
X
kxk k (j i)
j1
X
1
q
k=i
! p1
kxk kp
k=i
1
(j i) q
n1
X
! p1
kxk kp
k=1
for 1 i < j n.
Squaring the previous inequality, we get
kxj k2 + kxi k2 2 Re hxi , xj i + (j i)
2
q
n1
X
! p2
kxk kp
k=1
for 1 i < j n.
Utilising the same argument from the proof of Proposition 40, we
deduce the desired inequality (3.84).
Now assume that (3.84) holds with a constant E > 0, i.e.,

!2 n
! p2
n
n1
X 2
X
X
X
2

p
kxi k
xi + E
(j i) q
kxk k
,

i=1
i=1
1i<jn
k=1
for n 2 and xi H, i {1, . . . , n} , p > 1, p1 + 1q = 1.

For n = 2, x1 = 12 e, x2 = 12 e, kek = 1, we get 1 E, showing the
fact that the inequality (3.84) is sharp.
The particular case p = q = 2 is of interest [4].
Corollary 27. Let (H; h, i) be an inner product space and xi
H, i {1, . . . , n} . Then we have the inequality:

!2 n
n
n1
X 2 (n2 1) n X
X

(3.85)
kxi k
xi +
kxk k2 .

6
i=1
i=1
k=1
The constant
1
6
128
Proof. For p = q = 2, Proposition 41 provides the inequality

!2 n
n
n1
X 2
X
X
X

(j i)
kxk k2 ,
kxi k
xi +

i=1
i=1
and since
X
1i<jn
k=1
(j i)
1i<jn
= 1 + (1 + 2) + (1 + 2 + 3) + + (1 + 2 + + n 1)
=
n1
X
(1 + 2 + + k) =
k=1
n1
X
k (k + 1)
k=1
n (n2 1)
,
6
hence the inequality (3.84) is proved. The best constant may be shown
in the same way as above but we omit the details.
3.6.3. A Different Quadratic Inequality. Finally, we may state
and prove the following different result [4].
space, yi H, i {1, . . . , n} and M m > 0 are such that either
(3.86)
Re hM yj yi , yi myj i 0 for 1 i < j n,
or, equivalently,

1
M
+
m
yi
(M m) kyj k for 1 i < j n.
(3.87)
y
j

2
2
!2 n 2
n
n1
X
X
1 (M m)2 X

(3.88)
kyi k
yi +
k kyk+1 k2 .

2
M + m k=1
i=1
i=1
1 (M m)2
(3.89)
kyi k kyj k Re hyi , yj i =
kyj k2
4
M +m
kyi k2 +
(M m)2
kyj k2
M +m

M +m
2 Re yi ,
yj
2

+
1
(M m)2 kyj k2
n
3.7. FURTHER QUADRATIC REFINEMENTS
129
for 1 i < j n, and since, obviously,

(M m)2
M +m
kyi k kyj k kyi k2 +
kyj k2 ,
2
2
M +m
hence

2
M +m
2

kyi k kyj k

M +m
1
2 Re yi ,
yj + (M m)2 kyj k2 ,
2
n
giving the much simpler inequality

1 (M m)2
(3.90)
kyi k kyj k Re hyi , yj i
kyj k2 ,
4
M +m
for 1 i < j n.
m)2
Applying Lemma 5 for kij := 41 (M
kyj k2 , we deduce
M +m
!2 n 2
n
X
X
1 (M m)2 X

(3.91)
kyj k2
kyi k
yi +

2
M
+
m
i=1
i=1
1i<jn
with equality if and only if (3.90) holds for each i, j with 1 i < j n.
Since
X
X
X
X
kyj k2 =
kyj k2 +
kyj k2 + +
kyj k2
1i<jn
1<jn
n
X
n
X
j=2
j=3
kyj k2 +
n
X
j=2
2<jn
n1<jn
kyj k2 + +
(j 1) kyj k2 =
n
X
kyj k2 + kyn k2
j=n1
n1
X
k kyk+1 k2 ,
k=1
hence the inequality (3.88) is obtained.

3.7. Further Quadratic Refinements
3.7.1. The General Case. The following lemma is of interest in
itself as well [4].
over the real or complex number field K, xi H, i {1, . . . , n} and
k 1 with the property that:
(3.92)
kxi k kxj k k Re hxi , xj i ,
130
for each i, j with 1 i < j n. Then

n

!2
n
n
X 2
X
X

2
(3.93)
kxi k + (k 1)
kxi k k
xi .

i=1
i=1
i=1
The equality holds in (3.93) if and only if it holds in (3.92) for each
i, j with 1 i < j n.
Proof. Firstly, let us observe that the following identity holds
true:
n

!2
n
X 2
X

(3.94)
kxi k k
xi

i=1
i=1
*
+
n
n
n
X
X
X
=
kxi k kxj k k
xi ,
xj
=
i,j=1
n
X
i=1
j=1
[kxi k kxj k k Re hxi , xj i]
i,j=1
=2
[kxi k kxj k k Re hxi , xj i] + (1 k)
n
X
kxi k2 ,
i=1
1i<jn
since, obviously, Re hxi , xj i = Re hxj , xi i for any i, j {1, . . . , n} .

Using the assumption (3.92), we obtain
X
[kxi k kxj k k Re hxi , xj i] 0
1i<jn
and thus, from (3.94), we deduce the desired inequality (3.93).

The case of equality is obvious by the identity (3.94) and we omit
the details.
Remark 45. The inequality (3.93) provides the following reverse
of the quadratic generalised triangle inequality:

2
!2
n
n
n
n
X

X
X
X

(3.95) 0
kxi k
kxi k2 k
xi
kxi k2 .

i=1
i=1
i=1
i=1
P
Remark 46. Since k = 1 and ni=1 kxi k2 0, hence by (3.93) one
may deduce the following reverse of the triangle inequality

n
n

X
X

(3.96)
kxi k k
xi ,

i=1
i=1
provided (3.92) holds true for 1 i < j n.
131
The following corollary providing a better bound for

holds [4].
Pn
i=1
kxi k ,
Corollary 28. With the assumptions in Lemma 6, one has the

inequality:

r
n
n
X

X
nk

x
(3.97)
kxi k

i .

n + k 1 i=1
i=1
Proof. Using the Cauchy-Bunyakovsky-Schwarz inequality
!2
n
n
X
X
n
kxi k2
kxi k
i=1
i=1
we get
(3.98) (k 1)
n
X
kxi k +
i=1
n
X
!2
kxi k
i=1
!2
X
n
k1
+1
kxi k .
n
i=1
Consequently, by (3.98) and (3.93) we deduce

2
!2
n
n
X

X
n
+
k
1

k
xi
kxi k

n
i=1
i=1
giving the desired inequality (3.97).

3.7.2. Asymmetric Assumptions. The following result may be
stated as well [4].
space and xi H\ {0} , i {1, . . . , n} , (0, 1) , such that

x
j
xi
for 1 i < j n.
(3.99)

kxj k
!2
n
n

X
X
p
p
2
2
(3.100)
1
kxi k + 1 1
kxi k2
i=1
i=1

2
n
X

xi .

i=1
The case of equality holds in (3.100) iff

1
(3.101)
kxi k kxj k = p
Re hxi , xj i
1 2
for any 1 i < j n.
132
Proof. The condition (3.92) is obviously equivalent to

xj
2
2
kxi k + 1 2 Re xi ,
kxj k
for each 1 i <p
j n.
(3.102)

p
kxi k2
2
x
j
p
+ 1 2 p
Re xi ,
,
kxj k
1 2
1 2
for 1 i < j n.
p
(3.103)
+ q 2 pq, p, q 0, > 0
we have
p
kxi k2
(3.104)
2 kxi k p
+ 1 2 .
1 2
Making use of (3.102) and (3.104), we deduce that
1
kxi k kxj k p
Re hxi , xj i
1 2
for 1 i < j n.
Now, applying Lemma 5 for k = 1
12
, we deduce the desired
result.
Remark 47. If we assume that kxi k = 1, i {1, . . . , n} , satisfying
the simpler condition
(3.105)
kxj xi k
for 1 i < j n,
P
then, from (3.100), we deduce the following lower bound for k ni=1 xi k ,
namely

n

h
i 12 X
p

(3.106)
n + n (n 1) 1 2
xi .

i=1
p
The equality holds in (3.106) iff 1 2 = Re hxi , xj i for 1 i < j
n.
Remark 48. Under the hypothesis of Proposition 41, we have the
coarser but simpler reverse of the triangle inequality

n
n
X

X
p

4
(3.107)
1 2
kxi k
xi .

i=1
i=1
Also, applying Corollary 28 for k = 1
12
(3.108)
n
X
s
kxi k
i=1
133
, we can state that

n
X

p
p
xi ,

n 1 2 + 1 1 2 i=1
n
provided xi H satisfy (3.99) for 1 i < j n.

In the same manner, we can state and prove the following reverse
of the quadratic generalised triangle inequality [4].
space over the real or complex number field K, xi H, i {1, . . . , n}
and M m > 0 such that either
(3.109)
Re hM xj xi , xi mxj i 0 for 1 i < j n,
or, equivalently,

1
M
+
m
xi
(M m) kxj k for 1 i < j n
(3.110)
x
j

2
2
hold. Then
2 mM
(3.111)
M +m
n
X
i=1
!2
kxi k
2
n
M m X
M +m
kxi k2
i=1

2
n
X

xi .

i=1

(3.112)
M +m
kxi k kxj k =
Re hxi , xj i for 1 i < j n.
2 mM
Proof. From (3.109), observe that

kxi k2 + M m kxj k2 (M + m) Re hxi , xj i ,
for 1 i < j n. Dividing (3.113) by mM > 0, we deduce

(3.113)
M +m
kxi k2
Re hxi , xj i ,
+ mM kxj k2
mM
mM
and since, obviously
kxi k2
+ mM kxj k2
2 kxi k kxj k
mM
134
hence
M +m
kxi k kxj k
Re hxi , xj i , for 1 i < j n.
2 mM
Applying Lemma 6 for k =
M
+m
2 mM
1, we deduce the desired result.
Remark 49. We also must note that a simpler but coarser inequality that can be obtained from (3.111) is

! 12 n
n
X

X
2 mM

kxi k
xi ,

M +m
i=1
i=1
provided (3.109) holds true.

Finally, a different result related to the generalised triangle inequality is incorporated in the following theorem [4].
space over K, > 0 and xi H, i {1, . . . , n} with the property that
kxj xi k < kxj k for each i, j {1, . . . , n} .
(3.114)
Then we have the following reverse of the triangle inequality

Pn q
P
kxi k2 2
k ni=1 xi k
i=1
P
(3.115)
Pn
.
k ni=1 xi k
i=1 kxi k
The equality holds in (3.115) iff
q
(3.116) kxi k kxj k2 2 = Re hxi , xj i for each i, j {1, . . . , n} .
Proof. From (3.114), we have
kxi k2 + kxj k2 2 2 Re hxi , xj i ,
i, j {1, . . . , n} .
On the other hand,

q
2 kxi k kxj k2 2 kxi k2 + kxj k2 2 ,
i, j {1, . . . , n}
and thus
kxi k
kxj k2 2 Re hxi , xj i ,
i, j {1, . . . , n} .
Summing over i, j {1, . . . , n} , we deduce the desired inequality

(3.115).
The case of equality is also obvious from the above, and we omit
the details.
3.8. REVERSES FOR COMPLEX SPACES
135
3.8. Reverses for Complex Spaces

3.8.1. The Case of One Vector. The following result holds [5].
Theorem 54 (Dragomir, 2004). Let (H; h, i) be a complex inner
product space. Suppose that the vectors xk H, k {1, . . . , n} satisfy
the condition
(3.117)
0 r1 kxk k Re hxk , ei ,
0 r2 kxk k Im hxk , ei
for each k {1, . . . , n} , where e H is such that kek = 1 and r1 , r2

0. Then we have the inequality

n
n
q
X

X

2
2
(3.118)
r1 + r2
kxk k
xk ,

k=1
k=1

(3.119)
n
X
xk = (r1 + ir2 )
k=1
n
X
!
kxk k e.
k=1
Proof. In view of the Schwarz inequality in the complex inner

product space (H; h, i) , we have

2
2
*
+2
n
n
n
X

X

X

2
xk =
xk kek
(3.120)
xk , e

k=1
k=1
k=1
*
+
n
X
2

=
xk , e

k=1
n
!2
n
X

X

=
Re hxk , ei + i
Im hxk , ei

k=1
k=1
!2
!2
n
n
X
X
=
Re hxk , ei +
Im hxk , ei .
k=1
k=1
Now, by hypothesis (3.117)

(3.121)
n
X
!2
Re hxk , ei
r12
k=1
n
X
!2
kxk k
k=1
and
(3.122)
n
X
k=1
!2
Im hxk , ei
r22
n
X
k=1
!2
kxk k
136
If we add (3.121) and (3.122) and use (3.120), then we deduce the
desired inequality (3.118).
Now, if (3.119) holds, then

!
n
n
n
q
X

X
X

2
2
xk = |r1 + ir2 |
kxk k kek = r1 + r2
kxk k

k=1
k=1
k=1
and the case of equality is valid in (3.118).

Before we prove the reverse implication, let us observe that for
x H and e H, kek = 1, the following identity is true
kx hx, ei ek2 = kxk2 |hx, ei|2 ,
therefore kxk = |hx, ei| if and only if x = hx, ei e.
If we assume that equality holds in (3.118), then the case of equality
must hold in all the inequalities required in the argument used to prove
the inequality (3.118), and we may state that

*
+
n
n
X
X

(3.123)
xk =
xk , e ,

k=1
k=1
and
(3.124)
r1 kxk k = Re hxk , ei ,
r2 kxk k = Im hxk , ei
for each k {1, . . . , n} .

From (3.123) we deduce
(3.125)
n
X
xk =
k=1
* n
X
+
xk , e e
k=1
and from (3.124), by multiplying the second equation with i and summing both equations over k from 1 to n, we deduce
* n
+
n
X
X
(3.126)
(r1 + ir2 )
kxk k =
xk , e .
k=1
k=1
Finally, by (3.126) and (3.125), we get the desired equality (3.119).

The following corollary is of interest [5].
Corollary 29. Let e a unit vector in the complex inner product
space (H; h, i) and 1 , 2 (0, 1) . If xk H, k {1, . . . , n} are such
that
(3.127) kxk ek 1 ,
kxk iek 2
for each k {1, . . . , n} ,
137

n
n
q
X

X

2
2
(3.128)
2 1 2
kxk k
xk ,

k=1
k=1

!
q
X
n
n
q
X
(3.129)
xk =
1 21 + i 1 22
kxk k e.
k=1
k=1
Proof. From the first inequality in (3.127) we deduce

q
(3.130)
0 1 21 kxk k Re hxk , ei
for each k {1, . . . , n} .
From the second inequality in (3.127) we deduce
q
0 1 22 kxk k Re hxk , iei
for each k {1, . . . , n} . Since
Re hxk , iei = Im hxk , ei ,
hence
(3.131)
1 22 kxk k Im hxk , ei
for each k {1, . . . , n} .

Now, observe from (3.130) and (3.131),
that thep
condition (3.117)
p
2
of Theorem 54 is satisfied for r1 = 1 1 , r2 = 1 22 (0, 1) ,
and thus the corollary is proved.
The following corollary may be stated as well [5].
Corollary 30. Let e be a unit vector in the complex inner product
space (H; h, i) and M1 m1 > 0, M2 m2 > 0. If xk H, k
{1, . . . , n} are such that either
(3.132)
Re hM1 e xk , xk m1 ei 0,
Re hM2 ie xk , xk m2 iei 0
or, equivalently,
(3.133)

M
+
m
1
1
xk

e

2

xk M2 + m2 ie

2
1
(M1 m1 ) ,
2
1
(M2 m2 ) ,
2
138
for each k {1, . . . , n} , then we have the inequality

12 X
n
n

X
m 1 M1
m 2 M2

(3.134)
2
kxk k
xk .
2 +
2

(M1 + m1 )
(M2 + m2 )
k=1
k=1
!
X

n
n
X
m 2 M2
m 1 M1
kxk k e.
+i
(3.135)
xk = 2
M
M
2 + m2
1 + m1
k=1
k=1
Proof. From the first inequality in (3.132)
2 m 1 M1
(3.136)
0
kxk k Re hxk , ei
M1 + m 1
for each k {1, . . . , n} .
Now, the proof follows the same path as the one of Corollary 29
and we omit the details.
3.8.2. The Case of m Orthonormal Vectors. In [1], the authors have proved the following reverse of the generalised triangle inequality in terms of orthonormal vectors [5].
Theorem 55 (Diaz-Metcalf, 1966). Let e1 , . . . , em be orthonormal
vectors in (H; h, i), i.e., we recall that hei , ej i = 0 if i 6= j and kei k =
1, i, j {1, . . . , m} . Suppose that the vectors x1 , . . . , xn H satisfy
0 rk kxj k Re hxj , ek i ,
j {1, . . . , n} , k {1, . . . , m} . Then

! 21 n
n
m
X

X
X

(3.137)
kxj k
xj ,
rk2

k=1
j=1
j=1

(3.138)
n
X
j=1
xj =
n
X
!
kxj k
j=1
m
X
rk ek .
k=1
If the space (H; h, i) is complex and more information is available

for the imaginary part, then the following result may be stated as well
[5].
Theorem 56 (Dragomir, 2004). Let e1 , . . . , em H be an orthonormal family of vectors in the complex inner product space H. If
the vectors x1 , . . . , xn H satisfy the conditions
(3.139)
0 rk kxj k Re hxj , ek i ,
0 k kxj k Im hxj , ek i
139
for each j {1, . . . , n} and k {1, . . . , m} , then we have the following

reverse of the generalised triangle inequality;
n

" m
# 21 n
X
X
X

kxj k
xj .
(3.140)
rk2 + 2k

j=1
k=1
j=1

! m
n
n
X
X
X
(3.141)
xj =
kxj k
(rk + ik ) ek .
j=1
j=1
k=1
Proof. Before we prove the theorem, let us recall that, if x H

and e1 , . . . , em are orthogonal vectors, then the following identity holds
true:

2
m
n

X
X

2
(3.142)
hx, ek i ek = kxk
|hx, ek i|2 .
x

k=1
k=1
As a consequence of this identity, we note the Bessel inequality

(3.143)
m
X
|hx, ek i|2 kxk2 , x H.
k=1
The case of equality holds in (3.143) if and only if (see (3.142))

(3.144)
x=
m
X
hx, ek i ek .
k=1
P
Applying Bessels inequality for x = nj=1 xj , we have
n

* n
n
2
+2
m X
m X
X 2 X

X

(3.145)
xj
xj , ek =
hxj , ek i

j=1
j=1
k=1
k=1 j=1

!
!2
m X
n
n

X
X

=
Re hxj , ek i + i
Im hxj , ek i

j=1
j=1
k=1
!2
!2
m
n
n
X
X
X
=
Re hxj , ek i +
Im hxj , ek i .
k=1
j=1
Now, by the hypothesis (3.139) we have

!2
n
X
(3.146)
Re hxj , ek i rk2
j=1
j=1
n
X
j=1
!2
kxj k
140
and
(3.147)
n
X
!2
n
X
2k
Im hxj , ek i
j=1
!2
kxj k
j=1
Further, on making use of (3.145) (3.147), we deduce

2
!2
!2
n
n
n
m
X

X
X
X

rk2
xj
kxj k + 2k
kxj k

j=1
j=1
j=1
k=1
!
2 m
n
X
X

=
kxj k
rk2 + 2k ,
j=1
k=1

Now, if (3.141) holds, then

2
2
!2 m
n
n
X

X

X

xj =
kxj k
(rk + ik ) ek

j=1
j=1
k=1
!2 m
n
X
X
=
kxj k
|rk + ik |2
j=1
n
X
k=1
!2
kxj k
j=1
m
X

rk2 + 2k ,
k=1
and the case of equality holds in (3.140).

Conversely, if the equality holds in (3.140), then it must hold in all
the inequalities used to prove (3.140) and therefore we must have
n

n
2
m X
X 2 X

xj =
hxj , ek i
(3.148)

j=1
k=1
j=1
and
(3.149)
rk kxj k = Re hxj , ek i ,
k kxj k = Im hxj , ek i
for each j {1, . . . , n} and k {1, . . . , m} .

Using the identity (3.142), we deduce from (3.148) that
* n
+
n
m
X
X
X
(3.150)
xj =
xj , ek ek .
j=1
k=1
j=1
141
Multiplying the second equality in (3.149) with the imaginary unit i

and summing the equality over j from 1 to n, we deduce
* n
+
n
X
X
(3.151)
(rk + ik )
kxj k =
xj , ek
j=1
j=1
for each k {1, . . . , n} .

Finally, utilising (3.150) and (3.151), we deduce (3.141) and the
theorem is proved.
The following corollaries are of interest [5].
Corollary 31. Let e1 , . . . , em be orthonormal vectors in the complex inner product space (H; h, i) and k , k (0, 1) , k {1, . . . , n} .
If x1 , . . . , xn H are such that
kxj ek k k ,
kxj iek k k
for each j {1, . . . , n} and k {1, . . . , m} , then we have the inequality

# 21 n
" m
n
X

X
X

(3.152)
2 2k 2k
kxj k
xj .

j=1
k=1
j=1

! m

n
n
q
X
X
X q
2
2
(3.153)
xj =
kxj k
1 k + i 1 k ek .
j=1
j=1
k=1
The proof employs Theorem 56 and is similar to the one from Corollary 29. We omit the details.
Corollary 32. Let e1 , . . . , em be as in Corollary 31 and Mk
mk > 0, Nk nk > 0, k {1, . . . , m} . If x1 , . . . , xn H are such that
either
Re hMk ek xj , xj mk ek i 0, Re hNk iek xj , xj nk iek i 0
or, equivalently,

M
+
m
k
k
xj
ek

2

N
+
n
k
k
xj

ie
k

2
1
(Mk mk ) ,
2
1
(Nk nk )
2
for each j {1, . . . , n} and k {1, . . . , m} , then we have the inequality

( m
) 12 X
n
n
X

X
m k Mk
n k Nk

(3.154) 2
kxj k
xj .
2 +
2

(Mk + mk )
(Nk + nk )
j=1
j=1
k=1
142

! m

n
n
X
X
X
n k Nk
m k Mk
ek .
+i
(3.155)
xj = 2
kxj k
Nk + n k
Mk + m k
j=1
j=1
k=1
The proof employs Theorem 56 and is similar to the one in Corollary
30. We omit the details.
3.9. Applications for Vector-Valued Integral Inequalities
Let (H; h, i) be a Hilbert space over the real or complex number
field, [a, b] a compact interval in R and : [a, b] [0, ) a Lebesgue
Rb
integrable function on [a, b] with the property that a (t) dt = 1. If,
by L ([a, b] ; H) we denote the Hilbert space of all Bochner measurable
Rb
functions f : [a, b] H with the property that a (t) kf (t)k2 dt < ,
then the norm kk of this space is generated by the inner product
h, i : H H K defined by
Z b
hf, gi :=
(t) hf (t) , g (t)i dt.
a
The following proposition providing a reverse of the integral generalised

triangle inequality may be stated [3].
Proposition 42. Let (H; h, i) be a Hilbert space and : [a, b]
Rb
[0, ) as above. If g L ([a, b] ; H) is so that a (t) kg (t)k2 dt = 1
and fi L ([a, b] ; H) , i {1, . . . , n} , (0, 1) are so that
kfi (t) g (t)k
(3.156)
for a.e. t [a, b] and each i {1, . . . , n} , then we have the inequality
(3.157)
12
n Z b
X
p
2
2
1
(t) kfi (t)k dt
i=1

2 12
n
X

(t)
fi (t) dt .

i=1

21
n
n Z b
X
X
p
2
2
g (t)
fi (t) = 1
(t) kfi (t)k dt
i=1
for a.e. t [a, b] .
i=1
3.9. APPLICATIONS FOR VECTOR-VALUED INTEGRAL INEQUALITIES 143
Proof. Observe, by (3.157), that

Z b
12
2
kfi gk =
(t) kfi (t) g (t)k dt
a
Z
12
(t) 2 dt
=
for each i {1, . . . , n} . Applying Theorem 42 for the Hilbert space

L ([a, b] ; H) , we deduce the desired result.
Proposition 43. Let H, , g be as in Proposition 42. If fi
L ([a, b] ; H) , i {1, . . . , n} and M m > 0 are so that either
Re hM g (t) fi (t) , fi (t) mg (t)i 0
or, equivalently,

1
m
+
M
fi (t)
(M m)
g
(t)

2
2
for a.e. t [a, b] and each i {1, . . . , n} , then we have the inequality
(3.158)
12
n Z b
2 mM X
2
(t) kfi (t)k dt
m + M i=1
a

2 12
Z b
n
X

(t)
fi (t) dt .

a
i=1
12
n
n Z b
X
2 mM X
2
(t) kfi (t)k dt
g (t) ,
fi (t) =
m + M i=1
a
i=1
for a.e. t [a, b] .
The following proposition providing a reverse of the integral generalised triangle inequality may be stated [4].
Proposition 44. Let (H; h, i) be a Hilbert space and : [a, b]
Rb
[0, ) as above. If g L ([a, b] ; H) is so that a (t) kg (t)k2 dt = 1
and fi L ([a, b] ; H) , i {1, . . . , n} , and M m > 0 are so that
either
(3.159)
Re hM fj (t) fi (t) , fi (t) mfj (t)i 0
144
or, equivalently,

fi (t) m + M fj (t) 1 (M m) kfj (t)k

2
2
for a.e. t [a, b] and 1 i < j n, then we have the inequality
"
(3.160)
n Z
X
i=1
12 #2
(t) kfi (t)k2 dt

2
n
X

(t)
fi (t) dt

a
i=1
!
Z
n1
X
1 (M m)2 b
(t)
k kfk+1 (t)k2 dt.
+
2
m+M
a
k=1
Z

Z
21 Z b
12
(t) kfi (t)k2 dt
(t) kfj (t)k2 dt
(t) Re hfi (t) , fj (t)i dt

a
1 (M m)2
=
4
m+M
(t) kfj (t)k2 dt

Proof. We observe that
Re hM fj fi , fi mfj i
Z b
=
(t) Re hM fj (t) fi (t) , fi (t) mfj (t)i dt 0
a

Applying Theorem 50 for the Hilbert space L ([a, b] ; H) and for
yi = fi , i {1, . . . , n} , we deduce the desired result.
Another integral inequality incorporated in the following proposition holds [4]:
3.10. APPLICATIONS FOR COMPLEX NUMBERS
145
Proposition 45. With the assumptions of Proposition 44, we have

" n Z
21 #2
b
2 mM X
(3.161)
(t) kfi (t)k2 dt
m + M i=1
a

2
n Z b
M m X
+
(t) kfi (t)k2 dt
m+M
i=1 a

2
Z b
n
X

(t)
fi (t) dt.

a
i=1

Z
21 Z b
12
(t) kfi (t)k2 dt
(t) kfj (t)k2 dt
M +m
=
2 mM
(t) Re hfi (t) , fj (t)i dt

a

The proof is obvious by Theorem 52 and we omit the details.
3.10. Applications for Complex Numbers
The following reverse of the generalised triangle inequality with a
clear geometric meaning may be stated [5].
Proposition 46. Let z1 , . . . , zn be complex numbers with the property that
(3.162)
0 1 arg (zk ) 2 <
2
for each k {1, . . . , n} . Then we have the inequality

q
n
n
X

X

(3.163)
sin2 1 + cos2 2
|zk |
zk .

k=1
k=1

(3.164)
n
X
k=1
zk = (cos 2 + i sin 1 )
n
X
k=1
|zk | .
146
Proof. Let zk = ak + ibk . We may assume that bk 0, ak > 0,

k {1, . . . , n} , since, by (3.162), abkk = tan [arg (zk )] 0, 2 , k
{1, . . . , n} . By (3.162), we obviously have
0 tan2 1
b2k
tan2 2 ,
a2k
k {1, . . . , n}
from where we get

1
b2k + a2k
,
2
ak
cos2 2

k {1, . . . , n} , 2 0,
2

and
1 + tan2 1
1
a2k + b2k
=
,
2
2
ak
tan 1
sin2 1

k {1, . . . , n} , 1 0,
2

giving the inequalities

|zk | cos 2 Re (zk ) , |zk | sin 1 Im (zk )
for each k {1, . . . , n} .
Now, applying Theorem 54 for the complex inner product C endowed with the inner product hz, wi = z w for xk = zk , r1 = cos 2 ,
r2 = sin 1 and e = 1, we deduce the desired inequality (3.163). The
case of equality is also obvious by Theorem 54 and the proposition is
proven.
Another result that has an obvious geometrical interpretation is the
following one.
Proposition 47. Let c C with |z| = 1 and 1 , 2 (0, 1) . If
zk C, k {1, . . . , n} are such that
(3.165)
|zk c| 1 , |zk ic| 2 for each k {1, . . . , n} ,

n
n
q
X

X

2 21 22
|zk |
zk ,

(3.166)
k=1
k=1

!
q
X
n
n
q
X
1 21 + i 1 22
|zk | c.
(3.167)
zk =
k=1
k=1
The proof is obvious by Corollary 29 applied for H = C.
3.10. APPLICATIONS FOR COMPLEX NUMBERS
147
Remark 50. If we choose e = 1, and for 1 , 2 (0, 1) we define

(1, 1 ) := {z C| |z 1| 1 } , D
(i, 2 ) := {z C| |z i| 2 } ,
D
then obviously the intersection
(1, 1 ) D
(i, 2 )
S1 ,2 := D
is nonempty if and only if 1 + 2 2.

If zk S1 ,2 for k {1, . . . , n} , then (3.166) holds true. The
equality holds in (3.166) if and only if
q
X
n
n
q
X
2
2
zk =
1 1 + i 1 2
|zk | .
k=1
k=1
Bibliography
[1] J.B. DIAZ and F.T. METCALF, A complementary triangle inequality in
Hilbert and Banach spaces, Proceedings Amer. Math. Soc., 17(1) (1966), 88-97.
[2] S.S. DRAGOMIR, Advances in inequalities of the Schwarz,
Gruss and Bessel type in inner product spaces,
Preprint,
http://front.math.ucdavis.edu/math.FA/0309354.
[3] S.S. DRAGOMIR, Reverses of the triangle inequality in inner product
spaces, RGMIA Res. Rep. Coll., 7(2004), Supplement, Article 7, [ONLINE:
[4] S.S. DRAGOMIR, Quadratic reverses of the triangle inequality in inner product spaces, RGMIA Res. Rep. Coll., 7(2004), Supplement, Article 8, [ONLINE:
[5] S.S. DRAGOMIR, Some reverses of the generalised triangle inequality in complex inner product spaces, RGMIA Res. Rep. Coll., 7(2004), Supplement, Article 8, [ONLINE: http://rgmia.vu.edu.au/v7(E).html].
[6] J. KARAMATA, Teorija i Praksa Stieltjesova Integrala (Serbo-Coratian)
(Stieltjes Integral, Theory and Practice), SANU, Posebna izdanja, 154,
Beograd, 1949.
[7] S.M. KHALEELULA, On Diaz-Metcalfs complementary triangle inequality,
Kyungpook Math. J., 15 (1975), 9-11..
[8] M. MARDEN, The Geometry of the Zeros of a Polynomial in a Complex
Variable, Amer. Math. Soc. Math. Surveys, 3, New York, 1949.
C,
On a complementary inequality of the triangle inequality
[9] P.M. MILICI
(French), Mat. Vesnik 41(1989), No. 2, 83-88.
J.E. PECARI
and A.M. FINK, Classical and

C
New Inequalities in Analysis, Kluwer Academic Publishers, Dordrecht/Boston/London, 1993.
[11] M. PETROVICH, Module dune somme, L Ensignement Mathematique, 19
(1917), 53-56.
[12] H.S. WILF, Some applications of the inequality of arithmetic and geometric
means to polynomial equations, Proceedings Amer. Math. Soc., 14 (1963),
263-265.
149
CHAPTER 4
Reverses for the Continuous Triangle Inequality

4.1. Introduction
Let f : [a, b] K, K = C or R be a Lebesgue integrable function.
The following inequality, which is the continuous version of the triangle
inequality
Z b
Z b

(4.1)
f (x) dx
|f (x)| dx,

a
plays a fundamental role in Mathematical Analysis and its applications.

It appears, see [8, p. 492], that the first reverse inequality for (4.1)
was obtained by J. Karamata in his book from 1949, [6]. It can be
stated as
Z b

Z b

(4.2)
cos
|f (x)| dx
f (x) dx
a
provided
arg f (x) , x [a, b]
for given 0, 2 .
This integral inequality is the continuous version of a reverse inequality for the generalised triangle inequality
n
n
X
X

(4.3)
cos
|zi |
zi ,

i=1
i=1
provided
a arg (zi ) a + , for i {1, . . . , n} ,

where a R and 0, 2 , which, as pointed out in [8, p. 492],
was first discovered by M. Petrovich in 1917, [9], and, subsequently
rediscovered by other authors, including J. Karamata [6, p. 300 301],
H.S. Wilf [10], and in an equivalent form, by M. Marden [7].
The first to consider the problem in the more general case of Hilbert
and Banach spaces, were J.B. Diaz and F.T. Metcalf [1] who showed
151
152
4.
REVERSES FOR THE CONTINUOUS TRIANGLE INEQUALITY
that, in an inner product space H over the real or complex number

field , the following reverse of the triangle inequality holds

n
n
X

X

(4.4)
r
kxi k
xi ,

i=1
i=1
provided
Re hxi , ai
,
i {1, . . . , n} ,
kxi k
and a H is a unit vector, i.e., kak = 1. The case of equality holds in
(4.4) if and only if
!
n
n
X
X
(4.5)
xi = r
kxi k a.
0r
i=1
i=1
A generalisation of this result for orthonormal families is also well

known [1]:
Let a1 , . . . , am be m orthonormal vectors in H. Suppose the vectors
x1 , . . . , xn H\ {0} satisfy
0 rk
Re hxi , ak i
,
kxi k
Then
m
X
! 21
rk2
k=1
i {1, . . . , n} , k {1, . . . , m} .
n
X
i=1

n
X

kxi k
xi ,

i=1

n
X
i=1
xi =
n
X
i=1
!
kxi k
m
X
r k ak .
k=1
The main aim of this chapter is to survey some recent reverses of the
triangle inequality for Bochner integrable functions f with values in
Hilbert spaces and defined on a compact interval [a, b] R. Applications for Lebesgue integrable complex-valued functions are provided as
well.
4.2. Multiplicative Reverses
4.2.1. Reverses for a Unit Vector. We recall that f L ([a, b] ; H) ,
the space of Bochner integrable functions with values in a Hilbert space
H, if and only if f : [a, b] H is Bochner measurable on [a, b] and the
Rb
Lebesgue integral a kf (t)k dt is finite.
4.2. MULTIPLICATIVE REVERSES
153
Theorem 57 (Dragomir, 2004). If f L ([a, b] ; H) is such that

there exists a constant K 1 and a vector e H, kek = 1 with
(4.6)
kf (t)k K Re hf (t) , ei
for a.e. t [a, b] ,

Z b

Z b

(4.7)
kf (t)k dt K
f (t) dt

.
a

Z b

Z b
1
(4.8)
f (t) dt =
kf (t)k dt e.
K
a
a
Proof. By the Schwarz inequality in inner product spaces, we have
Z b
Z b

=
kek
(4.9)
f
(t)
dt
f
(t)
dt

a
a
Z b
Z b

f (t) dt, e Re
f (t) dt, e
a
a
Z b
Z b
Re
f (t) dt, e =
Re hf (t) , ei dt.
a
From the condition (4.6), on integrating over [a, b] , we deduce

Z
Z b
1 b
kf (t)k dt,
(4.10)
Re hf (t) , ei dt
K a
a
and thus, on making use of (4.9) and (4.10), we obtain the desired
inequality (4.7).
If (4.8) holds true, then, obviously
Z b

Z b
Z b

K
f (t) dt
kf (t)k dt =
kf (t)k dt,

= kek
a

If we assume that the equality holds in (4.7), then by the argument
provided at the beginning of our proof, we must have equality in each
of the inequalities from (4.9) and (4.10).
Observe that in Schwarzs inequality kxk kyk Re hx, yi , x, y H,
the case of equality holds if and only if there exists a positive scalar
such that
R b x = e. Therefore, equality holds in the first inequality in
(4.9) iff a f (t) dt = e, with 0 .
154
4.
If we assume that a strict inequality holds in (4.6) on a subset of

nonzero Lebesgue measure in [a, b] , then
Z b
Z b
kf (t)k dt < K
Re hf (t) , ei dt,
a
and by (4.9) we deduce a strict inequality in (4.7), which contradicts

the assumption. Thus, we must have kf (t)k = K Re hf (t) , ei for a.e.
t [a, b] .
If we integrate this equality, we deduce
Z b

Z b
Z b
kf (t)k dt = K
Re hf (t) , ei dt = K Re
f (t) dt, e
a
= K Re he, ei = K
giving
Z
1 b
kf (t)k dt,
=
K a
and thus the equality (4.8) is necessary.
This completes the proof.
A more appropriate result from an applications point of view is
perhaps the following result [2].
Corollary 33. Let e be a unit vector in the Hilbert space (H; h, i) ,
(0, 1) and f L ([a, b] ; H) so that
kf (t) ek
(4.11)
for a.e. t [a, b] .

Z b

Z b
p

,
1 2
kf (t)k dt
f
(t)
dt
(4.12)

a

Z b

Z b
p
(4.13)
f (t) dt = 1 2
kf (t)k dt e.
a

kf (t)k2 2 Re hf (t) , ei + 1 2 ,
giving
kf (t)k2 + 1 2 2 Re hf (t) , ei
for a.e. t [a, b]p
.
p
kf (t)k2
2 Re hf (t) , ei
p
(4.14)
+ 1 2 p
2
1
1 2
155
for a.e. t [a, b] .

p
+ q 2 pq, p, q 0, > 0
we have
p
kf (t)k2
(4.15)
2 kf (t)k p
+ 1 2
1 2
for each t [a, b] .
Making use of (4.14) and (4.15), we deduce
1
kf (t)k p
Re hf (t) , ei
1 2
for a.e. t [a, b] .
Applying Theorem 57 for K = 1
12
, we deduce the desired in-
equality (4.12).
In the same spirit, we also have the following corollary [2].
Corollary 34. Let e be a unit vector in H and M m > 0. If
f L ([a, b] ; H) is such that
(4.16)
Re hM e f (t) , f (t) mei 0
or, equivalently,
(4.17)

1

M
+
m
(M m)
f (t)
e
2

2
Z b

Z

2 mM b
(4.18)
kf (t)k dt
f (t) dt

,
M +m a
a
or, equivalently,
Z b

(0 )
kf (t)k dt
f (t) dt

a
a

2 Z

M m
b

.
f
(t)
dt

M +m
Z
(4.19)
The equality holds in (4.18) (or in the second part of (4.19)) if and
only if
Z b

Z b
2 mM
(4.20)
f (t) dt =
kf (t)k dt e.
M +m
a
a
156
4.
Proof. Firstly, we remark that if x, z, Z H, then the following

statements are equivalent
(i) Re hZ x, x zi 0
and
1
kZ zk .
(ii) x Z+z
2
2
Using this fact, we may simply realise that (4.14) and (4.15) are
equivalent.
Now, from (4.14), we obtain
kf (t)k2 + mM (M + m) Re hf (t) , ei
for a.e. t [a, b] . Dividing this inequality with mM > 0, we deduce

the following inequality that will be used in the sequel
kf (t)k2
M +m
(4.21)
+ mM
Re hf (t) , ei
mM
mM
for a.e. t [a, b] .
On the other hand
kf (t)k2
(4.22)
2 kf (t)k
+ mM ,
mM
for any t [a, b] .
Utilising (4.21) and (4.22), we may conclude with the following
inequality
M +m
kf (t)k
Re hf (t) , ei ,
2 mM
for a.e. t [a, b] .
Applying Theorem 57 for the constant K := 2m+M

1, we deduce
mM
the desired result.
4.2.2. Reverses for Orthonormal Families of Vectors. The
following result for orthonormal vectors in H holds [2].
Theorem 58 (Dragomir, 2004). Let {e1 , . . . , en } be a family of
orthonormal vectors in H, ki 0, i {1, . . . , n} and f L ([a, b] ; H)
such that
ki kf (t)k Re hf (t) , ei i
(4.23)
for each i {1, . . . , n} and for a.e. t [a, b] .

Then
! 21 Z
Z b

n
b
X

2

(4.24)
ki
kf (t)k dt
f (t) dt
,
i=1
157

Z b
X
Z b
n
(4.25)
ki ei .
f (t) dt =
kf (t)k dt
a
i=1
Rb
Proof. By Bessels inequality applied for a f (t) dt and the orthonormal vectors {e1 , . . . , en } , we have
Z b
2
2
n Z b
X

f
(t)
dt
(4.26)
f
(t)
dt,
e
i

a
i=1
n
X
Re
f (t) dt, ei
i=1
n Z b
X
i=1
2
Z
a
2
Re hf (t) , ei i dt .
Integrating (4.23), we get for each i {1, . . . , n}

Z b
Z b
0 ki
kf (t)k dt
Re hf (t) , ei i dt,
a
implying
(4.27)
n Z
X
i=1
2 X
2
Z b
n
2
Re hf (t) , ei i dt
ki
kf (t)k dt .
i=1
On making use of (4.26) and (4.27), we deduce

2
Z b
2
Z b
n
X

2

kf (t)k dt ,
ki
f (t) dt

a
i=1

If (4.25) holds true, then

Z b
Z b

n
X

f (t) dt =
kf (t)k dt
ki ei

a
a
i=1
# 21
"X
Z b
n
2
kf (t)k dt
ki2 kei k
=
a
n
X
i=1
i=1
ki2
! 12 Z
kf (t)k dt,
a
158
4.
Now, suppose that there is an i0 {1, . . . , n} for which

ki0 kf (t)k < Re hf (t) , ei0 i
on a subset of nonzero Lebesgue measure in [a, b] . Then obviously
Z b
Z b
ki0
kf (t)k dt <
Re hf (t) , ei0 i dt,
a
and using the argument given above, we deduce

! 12 Z
Z b

n
b
X

2
.
kf (t)k dt <
f
(t)
dt
ki

a
i=1
Therefore, if the equality holds in (4.24), we must have

ki kf (t)k = Re hf (t) , ei i
(4.28)
for each i {1, . . . , n} and a.e. t [a, b] .

Also, if the equality holds in (4.24), then we must have equality in
all inequalities (4.26), this means that

Z b
n Z b
X
(4.29)
f (t) dt =
f (t) dt, ei ei
a
i=1
and
Z
(4.30)
Im

= 0 for each i {1, . . . , n} .
f (t) dt, ei
a
Using (4.28) and (4.30) in (4.29), we deduce

Z b

Z b
n
X
f (t) dt =
Re
f (t) dt, ei ei
a
i=1
n Z b
X
i=1
Re hf (t) , ei i ei dt
n Z
X
i=1

kf (t)k dt ki ei
kf (t)k dt
=
a
n
X
ki ei ,
i=1
and the condition (4.25) is necessary.

This completes the proof.
The following two corollaries are of interest [2].
159
Corollary 35. Let {e1 , . . . , en } be a family of orthonormal vectors

in H, i (0, 1) , i {1, . . . , n} and f L ([a, b] ; H) such that:
(4.31)
kf (t) ei k i for i {1, . . . , n} and a.e. t [a, b] .

! 21 Z
Z b

n
b
X

2

kf (t)k dt
n
i
f (t) dt
,
a
i=1

Z b
Z b
n
X
1/2
f (t) dt =
kf (t)k dt
1 2i
ei .
a
i=1
Proof. From the proof of Theorem 57, we know that (4.25) implies
the inequality
q
1 2i kf (t)k Re hf (t) , ei i , i {1, . . . , n} , for a.e. t [a, b] .
p
Now, applying Theorem 58 for ki := 1 2i , i {1, . . . , n}, we deduce the desired result.
A different results is incorporated in (see [2]):
Corollary 36. Let {e1 , . . . , en } be a family of orthonormal vectors
in H, Mi mi > 0, i {1, . . . , n} and f L ([a, b] ; H) such that
(4.32)
Re hMi ei f (t) , f (t) mi ei i 0
or, equivalently,

1
M
+
m
i
i
f (t)
(Mi mi )
e
i

2
2
for i {1, . . . , n} and a.e. t [a, b] . Then we have the reverse of the
continuous triangle inequality
" n
# 12 Z
Z b

b
X 4mi Mi

,
kf
(t)k
dt
f
(t)
dt
2

(m
+
M
)
a
a
i
i
i=1
!
Z b
Z b
n
X
2 m i Mi
f (t) dt =
kf (t)k dt
ei .
m
+
M
i
i
a
a
i=1
160
4.
Proof. From the proof of Corollary 35, we know (4.32) implies

that
2 m i Mi
kf (t)k Re hf (t) , ei i , i {1, . . . , n} and a.e. t [a, b] .
m i + Mi
Now, applying Theorem 58 for ki :=
the desired result.
2 mi Mi
,
mi +Mi
i {1, . . . , n} , we deduce
4.3. Some Additive Reverses

4.3.1. The Case of a Unit Vector. The following result holds
[3].
Theorem 59 (Dragomir, 2004). If f L ([a, b] ; H) is such that
there exists a vector e H, kek = 1 and k : [a, b] [0, ), a Lebesgue
integrable function with
(4.33)
kf (t)k Re hf (t) , ei k (t)
for a.e. t [a, b] ,

Z b
Z b
Z b

(4.34)
(0 )
kf (t)k dt
f (t) dt
k (t) dt.

a

Z b
Z b
(4.35)
kf (t)k dt
k (t) dt
a
and
Z
(4.36)
Z
Z
kf (t)k dt
f (t) dt =
a

k (t) dt e.
Proof. If we integrate the inequality (4.33), we get

Z b
Z b
Z b
f (t) dt, e +
k (t) dt.
(4.37)
kf (t)k dt Re
a
Rb
By Schwarzs inequality for e and a f (t) dt, we have

Z b
(4.38)
Re
f (t) dt, e
a
Z b
Z b

f (t) dt, e
Re
f (t) dt, e
a
a
Z b

Z b

f (t) dt kek =
f (t) dt
.
a
4.3. SOME ADDITIVE REVERSES
161

(4.34).
If (4.35) and (4.36) hold true, then
Z b
Z b

Z b

=
kek
f
(t)
dt
kf
(t)k
dt
k
(t)
dt

a
a
b
a
b
kf (t)k dt
k (t) dt
and the equality holds true in (4.34).

valid and we need only to prove (4.36).
If kf (t)k Re hf (t) , ei < k (t) on a subset of nonzero Lebesgue
measure in [a, b] , then (4.37) holds as a strict inequality, implying that
(4.34) also holds as a strict inequality. Therefore, if we assume that
equality holds in (4.34), then we must have
kf (t)k = Re hf (t) , ei + k (t) for a.e. t [a, b] .
(4.39)
It is well known that in Schwarzs inequality kxk kyk Re hx, yi

the equality holds iff there exists a 0 such that x = y. Therefore,
if we assume that the equality holds in all of (4.38), then there exists
a 0 such that
Z b
(4.40)
f (t) dt = e.
a
Integrating (4.39) on [a, b] , we deduce

Z b
Z b
Z b
f (t) dt, e +
k (t) dt,
kf (t)k dt = Re
a
and thus, by (4.40), we get

Z b
Z b
2
kf (t)k dt = kek +
k (t) dt,
a
Rb
Rb
giving = a kf (t)k dt a k (t) dt.

Using (4.40), we deduce (4.36) and the theorem is completely proved.
The following corollary may be useful for applications [3].
Corollary 37. If f L ([a, b] ; H) is such that there exists a vector e H, kek = 1 and (0, 1) such that
(4.41)
kf (t) ek
for a.e. t [a, b] ,
162
4.

Z b

Z b

(4.42)
(0 )
kf (t)k dt
f (t) dt

a
a
Z b

2

Re
p
f (t) dt, e
p
2
2
a
1 1+ 1
Z b

2

f (t) dt
p

.
2
2
a
1 1+ 1
Z b

Z b
2

Re
f (t) dt, e
(4.43)
kf (t)k dt p
p
2
2
a
a
1 1+ 1
and
Z
(4.44)
f (t) dt
a
Z b
=
kf (t)k dt p
a
1+
Z
p
Re

f (t) dt, e e.
Proof. Firstly, note that (4.35) is equivalent to

kf (t)k2 + 1 2 2 Re hf (t) , ei ,
giving
p
kf (t)k2
2 Re hf (t) , ei
p
+ 1 2 p
1 2
1 2
for a.e. t [a, b] .
Since, obviously
p
kf (t)k2
2 kf (t)k p
+ 1 2
1 2
for any t [a, b] , then we deduce the inequality
kf (t)k
Re hf (t) , ei
p
1 2
for a.e. t [a, b] ,

2
kf (t)k Re hf (t) , ei p
1 2 1 +
1 2
Re hf (t) , ei
for a.e. t [a, b] .

Applying Theorem 59 for k (t) :=
12

1+ 12
163
Re hf (t) , ei , we
deduce the desired result.

In the same spirit, we also have the following corollary [3].
Corollary 38. If f L ([a, b] ; H) is such that there exists a vector e H, kek = 1 and M m > 0 such that either
Re hM e f (t) , f (t) mei 0
(4.45)
or, equivalently,
(4.46)

1

M
+
m
(M m)
f (t)
e
2

2

Z b

Z b

(4.47)
(0 )
kf (t)k dt
f (t) dt

a
2
Z b

M m
Re
f (t) dt, e
2 mM
a

2 Z

M m b

f (t) dt

.
2 mM
a

2
Z b

Z b
M m
kf (t)k dt
Re
f (t) dt, e
2 mM
a
a
and

2

Z
Z b
Z b
b
M m
Re
f (t) dt, e e.
f (t) dt =
kf (t)k dt
2 mM
a
a
a
Proof. Observe that (4.45) is clearly equivalent to
kf (t)k2 + mM (M + m) Re hf (t) , ei
for a.e. t [a, b] , giving the inequality
M +m
kf (t)k2
Re hf (t) , ei
+ mM
mM
mM
for a.e. t [a, b] .
164
4.
Since, obviously,
kf (t)k2
+ mM
2 kf (t)k
mM
for any t [a, b] , hence we deduce the inequality
M +m
Re hf (t) , ei for a.e. t [a, b] ,
kf (t)k
mM

2
M m
kf (t)k Re hf (t) , ei
Re hf (t) , ei
2 mM
for a.e. t [a, b] .
Finally, applying Theorem 59, we obtain the desired result.
We can state now (see also [3]):
Corollary 39. If f L ([a, b] ; H) and r L2 ([a, b] ; H) , e H,
kek = 1 are such that
(4.48)
kf (t) ek r (t)
for a.e. t [a, b] ,

Z b

Z
Z b

1 b 2

r (t) dt.
(4.49)
(0 )
kf (t)k dt
f (t) dt
2 a
a
a
Z
Z b
1 b 2
r (t) dt
kf (t)k dt
2 a
a
and
Z b

Z b
Z
1 b 2
f (t) dt =
kf (t)k dt
r (t) dt e.
2 a
a
a
Proof. The condition (4.48) is obviously equivalent to
kf (t)k2 + 1 2 Re hf (t) , ei + r2 (t)
for a.e. t [a, b] .
Using the elementary inequality
2 kf (t)k kf (t)k2 + 1, t [a, b] ,
we deduce
for a.e. t [a, b] .
1
kf (t)k Re hf (t) , ei r2 (t)
2
165
Applying Theorem 59 for k (t) := 12 r2 (t) , t [a, b], we deduce the

desired result.
Finally, we may state and prove the following result as well [3].
Corollary 40. If f L ([a, b] ; H), e H, kek = 1 and M, m :
m)2
[a, b] [0, ) with M m a.e. on [a, b] , are such that (M
M +m
L [a, b] and either

1

M
(t)
+
m
(t)
f (t)
[M (t) m (t)]
e
(4.50)

2
2
or, equivalently,
(4.51)
Re hM (t) e f (t) , f (t) m (t) ei 0

Z b

Z
Z b

1 b [M (t) m (t)]2

(4.52) (0 )
kf (t)k dt
f (t) dt
dt.
4 a M (t) + m (t)
a
a
Z b
Z
1 b [M (t) m (t)]2
kf (t)k dt
dt
4 a M (t) + m (t)
a
and
!
Z b
Z b
Z b
2
1
[M (t) m (t)]
f (t) dt =
kf (t)k dt
dt e.
4 a M (t) + m (t)
a
a
Proof. The condition (4.50) is equivalent to

2
M (t) + m (t)
2
kf (t)k +
2

M (t) + m (t)
1
2
Re hf (t) , ei + [M (t) m (t)]2
2
4
for a.e. t [a, b] , and since

2
M (t) + m (t)
M (t) + m (t)
2
2
kf (t)k kf (t)k +
, t [a, b]
2
2
hence
1 [M (t) m (t)]2
4 M (t) + m (t)
for a.e. t [a, b] .
(t)m(t)]2
Now, applying Theorem 59 for k (t) := 14 [M
, t [a, b], we
M (t)+m(t)
deduce the desired inequality.
166
4.
4.3.2. Additive Reverses for Orthonormal Families. The following reverse of the continuous triangle inequality for vector valued
integrals holds [3].
Theorem 60 (Dragomir, 2004). Let f L ([a, b] ; H) , where H is
a Hilbert space over the real or complex number field K, {ei }i{1,...,n}
an orthonormal family in H and Mi L [a, b] , i {1, . . . , n} . If we
assume that
kf (t)k Re hf (t) , ei i Mi (t) for a.e. t [a, b] ,
(4.53)

Z b

Z b
n Z b
1X
1
(4.54)
kf (t)k dt
f (t) dt
Mi (t) dt.
+ n
n a
a
a
i=1
Z b
n Z
1X b
Mi (t) dt
(4.55)
kf (t)k dt
n i=1 a
a
and
b
Z
(4.56)
Z
f (t) dt =
1X
kf (t)k dt
n i=1
Mi (t) dt
a
n
X
ei .
i=1
Proof. If we integrate the inequality (4.53) on [a, b] , we get

Z b
Z b
Z b
kf (t)k dt Re
f (t) dt, ei +
Mi (t) dt
a
for each i {1, . . . , n} . Summing these inequalities over i from 1 to n,

we deduce
*Z
+
Z b
n
n Z
b
X
1
1X b
f (t) dt,
ei +
Mi (t) dt.
(4.57)
kf (t)k dt Re
n
n i=1 a
a
a
i=1
Rb
P
By Schwarzs inequality for a f (t) dt and ni=1 ei , we have
*Z
+
n
b
X
(4.58)
Re
f (t) dt,
ei
a
i=1
*Z
+ *Z
+
n
n

b
b
X
X

Re
f (t) dt,
ei
f (t) dt,
ei

a

a
i=1
i=1

Z b

Z b
X

n

f (t) dt
ei = n
f (t) dt
,

a
a
i=1
167
since

v

2 v
u n
n
n
X
u
X

u
uX

t

ei =
ei = t
kei k2 = n.

i=1
i=1
i=1

(4.54).

Z
Z b

n Z b
n
X

b
X

1
1
1

kf (t)k dt
Mi (t) dt
ei
f (t) dt =

n i=1 a
n a
n a
i=1
!
Z b
n Z
1X b
=
kf (t)k dt
Mi (t) dt
n i=1 a
a
and the equality in (4.54) holds true.
Conversely, if the equality holds in (4.54), then, obviously, (4.55) is
valid.
Taking into account the argument presented above for the previous
result (4.54), it is obvious that, if the equality holds in (4.54), then it
must hold in (4.53) for a.e. t [a, b] and for each i {1, . . . , n} and
also the equality must hold in any of the inequalities in (4.58).
It is well known that in Schwarzs inequality Re hu, vi kuk kvk ,
the equality occurs if and only if u = v with 0, consequently, the
equality holds in all inequalities from (4.58) simultaneously iff there
exists a 0 with
Z b
n
X
(4.59)
ei =
f (t) dt.
a
i=1
If we integrate the equality in (4.53) and sum over i, we deduce

*Z
+
Z b
n
n Z b
b
X
X
(4.60) n
f (t) dt = Re
f (t) dt,
ei +
Mi (t) dt.
a
Replacing
(4.61)
Rb
a
i=1
i=1
f (t) dt from (4.59) into (4.60), we deduce

n 2
Z b
n Z b
X
X

n
f (t) dt =
ei +
Mi (t) dt

a
i=1
i=1 a
n Z b
X
= n +
Mi (t) dt.
i=1
Finally, we note that (4.59) and (4.61) will produce the required
identity (4.56), and the proof is complete.
168
4.
The following corollaries may be of interest for applications [3].

Corollary 41. Let f L ([a, b] ; H) , {ei }i{1,...,n} an orthonormal
family in H and i (0, 1) , i {1, . . . , n} such that
kf (t) ei k i for a.e. t [a, b] .
(4.62)
Then we have the inequalities:

Z b

Z b

1

(4.63)
kf (t)k dt
f
(t)
dt

n a
a
+
*Z
n
b
2i
1X

ei
+ Re
f (t) dt,
p
p
n i=1 1 2 1 + 1 2
a
i
i
Z b

1

f
(t)
dt

n a
12
n
2
1X
i p

1 +
.
p
n i=1 1 2 1 + 1 2
i
i
The equality holds in the first inequality in (4.63) if and only if
*Z
+
Z b
n
b
1X
2i

ei
kf (t)k dt Re
f (t) dt,
p
p
n i=1 1 2 1 + 1 2
a
a
i
and
Z
f (t) dt
a
*Z
kf (t)k dt Re
=
a
+
n
2
X
1
i p
ei
f (t) dt,
p
n i=1 1 2 1 + 1 2
i
n
X
ei .
i=1
Proof. As in the proof of Corollary 37, the assumption (4.62)

implies
2i

Re hf (t) , ei i
kf (t)k Re hf (t) , ei i p
p
2
2
1 i
1 i + 1
for a.e. t [a, b] and for each i {1, . . . , n} .
169
Now, if we apply Theorem 60 for

2 Re hf (t) , ei i
, i {1, . . . , n} , t [a, b] ,
Mi (t) := p i p
2
2
1 i
1 i + 1
we deduce the first inequality in (4.63).
By Schwarzs inequality in H, we have
+
*Z
n
b
2i
1X

ei
Re
f (t) dt,
p
p
n i=1 1 2 1 + 1 2
a
i
i

Z b
n

2
X

1

i

ei

f (t) dt
p
p

n

a
i=1 1 2 1 + 1 2
i
Z b

n
X

1
p
f (t) dt
=

n
a
i=1
2 12
2i
1
2i
1+
2i
,
which implies the second inequality in (4.63).

The second result is incorporated in [3]:
family in H and Mi mi > 0 such that either
(4.64)
Re hMi ei f (t) , f (t) mi ei i 0
or, equivalently,

1

M
+
m
i
i
(Mi mi )
f (t)
e
i
2

2
for a.e. t [a, b] and each i {1, . . . , n} .Then we have
Z b

Z b

1

(4.65)
kf (t)k dt
f (t) dt

n a
a
*Z
2 +
n
b
Mi m i
1X
+ Re
f (t) dt,
ei
n i=1
2 m i Mi
a
4 ! 12
Z b

n
X

Mi m i
1
1 + 1
.
f (t) dt

n i=1
4mi Mi
n a
The equality holds in the first inequality in (4.65) if and only if
*Z
2 +
Z b
n
b
Mi m i
1X
kf (t)k dt Re
f (t) dt,
ei
n i=1
2 m i Mi
a
a
170
4.
and
Z
f (t) dt
a
*Z
kf (t)k dt Re
=
a
1X
f (t) dt,
n i=1
2 +!
Mi m i
ei
2 m i Mi
n
X
ei .
i=1
Proof. As in the proof of Corollary 38, from (4.64), we have
2
Mi m i
kf (t)k Re hf (t) , ei i
Re hf (t) , ei i
2 m i Mi
for a.e. t [a, b] and i {1, . . . , n} .
Applying Theorem 60 for
2
Mi m i
Mi (t) :=
Re hf (t) , ei i ,
2 m i Mi
t [a, b] , i {1, . . . , n} ,

In a different direction, we may state the following result as well
[3].
family in H and ri L2 ([a, b]) , i {1, . . . , n} such that
kf (t) ei k ri (t)
for a.e. t [a, b] and i {1, . . . , n} .

Z b

Z b
n Z b
X

1
1
2
(4.66)
kf (t)k dt
f (t) dt
ri (t) dt .
+ 2n
n a
a
a
i=1

Z b
n Z b
1 X
2
kf (t)k dt
ri (t) dt
2n i=1
a
a
and
Z
"Z
f (t) dt =
a
1X
kf (t)k dt
n i=1
Z
a
# X
n
2
ri (t) dt
ei .
i=1
171
Proof. As in the proof of Corollary 39, from (4.48), we deduce

that
1
(4.67)
kf (t)k Re hf (t) , ei i ri2 (t)
2
for a.e. t [a, b] and i {1, . . . , n} .
1
Mi (t) := ri2 (t) , t [a, b] , i {1, . . . , n} ,
2
we get the desired result.
Finally, the following result holds [3].
family in H, Mi , mi : [a, b] [0, ) with Mi mi a.e. on [a, b] and
(Mi mi )2
L [a, b] , and either
Mi +mi

1
M
(t)
+
m
(t)
i
i
[Mi (t) mi (t)]2
(4.68)
f
(t)
e
i

2
2
or, equivalently,
Re hMi (t) ei f (t) , f (t) mi (t) ei i 0
for a.e. t [a, b] and any i {1, . . . , n}, then we have the inequality
Z b

Z b

1

(4.69)
kf (t)k dt
f
(t)
dt

n a
a
!
Z b
n
1 X
[Mi (t) mi (t)]2
+
dt .
4n i=1
Mi (t) + mi (t)
a
!
Z b
Z b
n
1 X
[Mi (t) mi (t)]2
kf (t)k dt
dt
4n i=1
Mi (t) + mi (t)
a
a
and
Z
f (t) dt
a
Z
=
a
1 X
kf (t)k dt
4n i=1
Z
a
!! n
X
[Mi (t) mi (t)]2
dt
ei .
Mi (t) + mi (t)
i=1
172
4.
Proof. As in the proof of Corollary 40, (4.68), implies that

kf (t)k Re hf (t) , ei i
1 [Mi (t) mi (t)]2
4 Mi (t) + mi (t)
for a.e. t [a, b] and i {1, . . . , n} .

Mi (t) :=
1 [Mi (t) mi (t)]2
,
4 Mi (t) + mi (t)
t [a, b] , i {1, . . . , n} ,

4.4.1. Additive Reverses. The following lemma holds [4].
Lemma 7 (Dragomir, 2004). Let f L ([a, b] ; H) be such that there
exists a function k : R2 R, := {(t, s) |a t s b} with
the property that k L () and
(0 ) kf (t)k kf (s)k Re hf (t) , f (s)i k (t, s) ,
(4.70)
for a.e. (t, s) . Then we have the following quadratic reverse of the
continuous triangle inequality:
2
Z b
2 Z b
ZZ

(4.71)
kf (t)k dt
f (t) dt + 2
k (t, s) dtds.
a
The case of equality holds in (4.71) if and only if it holds in (4.70) for
a.e. (t, s) .
Proof. We observe that the following identity holds
2
Z b
2 Z b

(4.72)
kf (t)k dt
f (t) dt

a
a

Z b
Z b
Z bZ b
=
kf (t)k kf (s)k dtds
f (t) dt,
f (s) ds
a
Z bZ
Z bZ
kf (t)k kf (s)k dtds
=
a
Z bZ
Re hf (t) , f (s)i dtds

a
[kf (t)k kf (s)k Re hf (t) , f (s)i] dtds := I.
=
a
Now, observe that for any (t, s) [a, b] [a, b] , we have

kf (t)k kf (s)k Re hf (t) , f (s)i
= kf (s)k kf (t)k Re hf (s) , f (t)i
173
and thus
ZZ
(4.73)
[kf (t)k kf (s)k Re hf (t) , f (s)i] dtds.
I=2
Using the assumption (4.70), we deduce

ZZ
ZZ
[kf (t)k kf (s)k Re hf (t) , f (s)i] dtds
k (t, s) dtds,
and, by the identities (4.72) and (4.73), we deduce the desired inequality (4.71).
The case of equality is obvious and we omit the details.
Remark 51. From (4.71) one may deduce a coarser inequality that
can be useful in some applications. It is as follows:
Z b

Z Z
12
Z b

(0 )
kf (t)k dt
k (t, s) dtds .
f (t) dt

2
a
Remark 52. If the condition (4.70) is replaced with the following

refinement of the Schwarz inequality
(4.74)
(0 ) k (t, s) kf (t)k kf (s)k Re hf (t) , f (s)i
for a.e. (t, s) , then the following refinement of the quadratic triangle inequality is valid
2
Z b
2 Z b
ZZ

(4.75)
kf (t)k dt
f (t) dt + 2
k (t, s) dtds
a
a
Z b
2 !

.

f
(t)
dt

a
The equality holds in (4.75) iff the case of equality holds in (4.74) for
a.e. (t, s) .
Theorem 61 (Dragomir, 2004). Let f L ([a, b] ; H) be such that
there exists M 1 m 0 such that either
(4.76)
Re hM f (s) f (t) , f (t) mf (s)i 0
or, equivalently,

1
M +m

(4.77)
f (s)
f (t)
2 (M m) kf (s)k
2
174
4.
for a.e. (t, s) . Then we have the inequality:

2
Z b
2 Z b

(4.78)
kf (t)k dt
f (t) dt

a
1 (M m)2
+
2
M +m
(s a) kf (s)k2 ds.

1 (M m)2
(4.79) kf (t)k kf (s)k Re hf (t) , f (s)i =
kf (s)k2
4
M +m
for a.e. (t, s) .

2
M +m
2
kf (t)k +
kf (s)k2
2

M +m
1
2 Re f (t) ,
f (s) + (M m)2 kf (s)k2 ,
2
4
for a.e. (t, s) , and obviously, since

2
M +m
M +m
2
2
kf (t)k kf (s)k kf (t)k +
kf (s)k2 ,
2
2
we deduce that

M +m
2
kf (t)k kf (s)k
2

M +m
1
2 Re f (t) ,
f (s) + (M m)2 kf (s)k2 ,
2
4
giving the much simpler inequality:
(4.80)
for a.e. (t, s) .

Applying Lemma 7 for k (t, s) :=
Z
(4.81)
a
1
4
1 (M m)2
kf (s)k2
4
M +m
(M m)2
M +m
kf (s)k2 , we deduce
2
2 Z b

kf (t)k dt
f
(t)
dt

a
1 (M m)2
+
2
M +m
ZZ
kf (s)k2 ds
with equality if and only if (4.80) holds for a.e. (t, s) .
175
Since

ZZ
Z b Z s
Z b
2
2
kf (s)k ds =
kf (s)k dt ds =
(s a) kf (s)k2 ds,
then by (4.81) we deduce the desired result (4.78).

Another result which is similar to the one above is incorporated in
the following theorem [4].
Theorem 62 (Dragomir, 2004). With the assumptions of Theorem
61, we have
Z
(4.82)
a
2
2 Z b

kf (t)k dt
f
(t)
dt

a
or, equivalently,
1

Z b
M +m 2
(4.83)
kf (t)k dt
2 Mm
a
2
M m
2 Mm
Z b
2

f
(t)
dt

a

Z b

.
f
(t)
dt

a
The case of equality holds in (4.82) or (4.83) if and only if

(4.84)
M +m
kf (t)k kf (s)k =
Re hf (t) , f (s)i ,
2 Mm
for a.e. (t, s) .

Proof. From (4.76), we deduce
kf (t)k2 + M m kf (s)k2 (M + m) Re hf (t) , f (s)i
for a.e. (t, s) . Dividing by M m > 0, we deduce

kf (t)k2
M +m
Re hf (t) , f (s)i
+ M m kf (s)k2
Mm
Mm
and, obviously, since
kf (t)k2
+ M m kf (s)k2 ,
2 kf (t)k kf (s)k
Mm
hence
M +m
Re hf (t) , f (s)i
kf (t)k kf (s)k
Mm
176
4.
for a.e. (t, s) , giving

Applying Lemma 7 for k (t, s) :=
2
M m
Re hf (t) , f (s)i .
2 Mm
2
M m)
Mm
Re hf (t) , f (s)i , we deduce
2
2 Z b

(4.85)
kf (t)k dt
f (t) dt

a
a

2
M m
+
Re hf (t) , f (s)i .
2 Mm
On the other hand, since
Z
Re hf (t) , f (s)i = Re hf (s) , f (t)i for any (t, s) [a, b]2 ,

hence
Z Z
1 b b
Re hf (t) , f (s)i dtds =
Re hf (t) , f (s)i dtds
2 a a
Z b

Z b
1
= Re
f (t) dt,
f (s) ds
2
a
a
Z b
2

1
f (t) dt
=

2 a
ZZ
and thus, from (4.85), we get (4.82).

The equivalence between (4.82) and (4.83) is obvious and we omit
the details.
4.4.2. Related Results. The following result also holds [4].
Theorem 63 (Dragomir, 2004). Let f L ([a, b] ; H) and , R
be such that either
(4.86)
Re hf (s) f (t) , f (t) f (s)i 0
or, equivalently,

1
f (t)
(4.87)
f (s)

2 | | kf (s)k
2
for a.e. (t, s) . Then we have the inequality:
Z b
+
(4.88)
[(b s) + (s a)] kf (s)k2 ds
2
a
Z b
2

f (s) ds

.
a
177
The case of equality holds in (4.88) if and only if the case of equality
holds in either (4.86) or (4.87) for a.e. (t, s) .
Proof. The inequality (4.86) is obviously equivalent to
kf (t)k2 + kf (s)k2 ( + ) Re hf (t) , f (s)i
(4.89)
for a.e. (t, s) .

Integrating (4.89) on , we deduce
Z b Z
(4.90)
a

Z b
Z s
2
dt ds
kf (t)k dt ds +
kf (s)k
a
a

Z b Z s
= ( + )
Re hf (t) , f (s)i dt ds.
2
It is easy to see, on integrating by parts, that

Z b Z
a
b Z b

Z s

2
kf (t)k dt
s kf (s)k2 ds
kf (t)k dt ds = s
2
a
s
a
b
=b
kf (s)k2 ds
s kf (s)k2 ds
a
a
Z b
(b s) kf (s)k2 ds
=
a
and
Z b
kf (s)k
a

Z b
(s a) kf (s)k2 ds.
dt ds =
Since
d
ds
Z b
2 !
Z s

Z s

d

f (t) dt
f (t) dt,
f (t) dt

= ds
a
a
a

Z s

Z s
= f (s) ,
f (t) dt +
f (t) dt, f (s)
a
a
Z s

= 2 Re
f (t) dt, f (s) ,
a
178
4.
hence
Z b Z
a

Z s

Z b
Re hf (t) , f (s)i dt ds =
Re
f (t) dt, f (s) ds
a
a
Z s
2 !
Z

1 b d

ds
=
f
(t)
dt

2 a ds a
Z b
2

1
.
=
f
(t)
dt

2 a
Utilising (4.90), we deduce the desired inequality (4.88).

The case of equality is obvious and we omit the details.
Remark 53. Consider the function (s) := (b s) + (s a) ,
s [a, b] . Obviously,
(s) = ( 1) s + b a.
Observe that, if 1, then
b a = (a) (s) (b) = (b a) ,
s [a, b]
and, if < 1, then

(b a) (s) b a,
s [a, b] .
Taking into account the above remark, we may state the following
corollary [4].
Corollary 45. Assume that f, , are as in Theorem 63.
a) If 1, then we have the inequality
Z b
2
Z b

+
2

.
(b a)
kf (s)k ds
f
(s)
ds

2
a
a
b) If 0 < < 1, then we have the inequality
Z b
2
Z b

+
2

.
(b a)
kf (s)k ds
f
(s)
ds

2 a
a
4.5. Refinements for Complex Spaces
4.5.1. The Case of a Unit Vector. The following result holds
[5].
Theorem 64 (Dragomir, 2004). Let (H; h, i) be a complex Hilbert
space. If f L ([a, b] ; H) is such that there exists k1 , k2 0 with
(4.91)
k1 kf (t)k Re hf (t) , ei , k2 kf (t)k Im hf (t) , ei
4.5. REFINEMENTS FOR COMPLEX SPACES
179
for a.e. t [a, b] , where e H, kek = 1, is given, then

Z b

Z b
q

2
2
.
(4.92)
k1 + k 2
kf (t)k dt
f
(t)
dt

a

Z b

Z b
(4.93)
f (t) dt = (k1 + ik2 )
kf (t)k dt e.
a
Proof. Using the Schwarz inequality kuk kvk |hu, vi| , u, v H;

in the complex Hilbert space (H; h, i) , we have
(4.94)
Z b
2 Z b
2

=
kek2
f
(t)
dt
f
(t)
dt

a
a
Z b
2
2 Z b

f (t) dt, e =
hf (t) , ei dt
a
Z b
Z b
2

=
Re hf (t) , ei dt + i
Im hf (t) , ei dt
a
a
Z b
2 Z b
2
=
Re hf (t) , ei dt +
Im hf (t) , ei dt .
a
Now, on integrating (4.91), we deduce

Z b
Z b
(4.95)
Re hf (t) , ei dt,
k1
kf (t)k dt
a
a
Z b
Z b
k2
kf (t)k dt
Im hf (t) , ei dt
a
implying
b
Z
(4.96)
a
2
2
Z b
2
Re hf (t) , ei dt k1
kf (t)k dt
a
and
Z
(4.97)
a
Z b
2
2
2
kf (t)k dt .
Im hf (t) , ei dt k2
a
If we add (4.96) and (4.97) and use (4.94), we deduce the desired inequality (4.92).
180
4.
Further, if (4.93) holds, then obviously

Z b

Z b

f (t) dt = |k1 + ik2 |
kf (t)k dt kek

a
k12 + k22
a
b
kf (t)k dt,
a
and the equality case holds in (4.92).

Before we prove the reverse implication, let us observe that, for
x H and e H, kek = 1, the following identity is valid
kx hx, ei ek2 = kxk2 |hx, ei|2 ,
therefore kxk = |hx, ei| if and only if x = hx, ei e.
If we assume that equality holds in (4.92), then the case of equality
must hold in all the inequalities required in the argument used to prove
the inequality (4.92). Therefore, we must have
Z b
Z b

=

(4.98)
f
(t)
dt
f
(t)
dt,
e

a
and
(4.99)
k1 kf (t)k = Re hf (t) , ei ,
k2 kf (t)k = Im hf (t) , ei
for a.e. t [a, b] .

From (4.98) we deduce

Z b
Z b
f (t) dt, e e,
f (t) dt =
(4.100)
a
and from (4.99), by multiplying the second equality with i, the imaginary unit, and integrating both equations on [a, b] , we deduce
Z b

Z b
f (t) dt, e .
kf (t)k dt =
(4.101)
(k1 + ik2 )
a
Finally, by (4.100) and (4.101), we deduce the desired equality (4.93).

The following corollary is of interest [5].
Corollary 46. Let e be a unit vector in the complex Hilbert space
(H; h, i) and 1 , 2 (0, 1) . If f L ([a, b] ; H) is such that
(4.102)
kf (t) ek 1 , kf (t) iek 2

Z b

Z b
q

2
2

(4.103)
2 1 2
kf (t)k dt
f (t) dt
.
a
181

q
Z b

Z b
q
2
2
(4.104)
f (t) dt =
1 1 + i 1 2
kf (t)k dt e.
a
Proof. From the first inequality in (4.102) we deduce, by taking

the square, that
kf (t)k2 + 1 21 2 Re hf (t) , ei ,
implying
(4.105)
kf (t)k2
p
+
1 21
1 21
2 Re hf (t) , ei
p
1 21
for a.e. t [a, b] .

Since, obviously
(4.106)
kf (t)k2
2 kf (t)k p
+
1 21
1 21 ,
hence, by (4.105) and (4.106) we get

q
(4.107)
0 1 21 kf (t)k Re hf (t) , ei
for a.e. t [a, b] .
From the second inequality in (4.102) we deduce
q
0 1 22 kf (t)k Re hf (t) , iei
for a.e. t [a, b] . Since
Re hf (t) , iei = Im hf (t) , ei
hence
(4.108)
q
1 22 kf (t)k Im hf (t) , ei
for a.e. t [a, b] .

Now, observe from (4.107) and
p(4.108), that the
p condition (4.91) of
2
Theorem 64 is satisfied for k1 = 1 1 , k2 = 1 22 (0, 1) , and
thus the corollary is proved.
The following corollary may be stated as well [5].
Corollary 47. Let e be a unit vector in the complex Hilbert space
(H; h, i) and M1 m1 > 0, M2 m2 > 0. If f L ([a, b] ; H) is such
that either
(4.109)
Re hM1 e f (t) , f (t) m1 ei 0,

Re hM2 ie f (t) , f (t) m2 iei 0
182
4.
or, equivalently,

f (t) M1 + m1 e

2

M
+
m
2
2
f (t)
ie

2
(4.110)
1
(M1 m1 ) ,
2
1
(M2 m2 ) ,
2

m 1 M1
m 2 M2
(4.111) 2
2 +
(M1 + m1 )
(M2 + m2 )2
12 Z
kf (t)k dt
a
Z b

.

f
(t)
dt

a

Z b

Z b
m 1 M1
m 2 M2
(4.112)
f (t) dt = 2
+i
kf (t)k dt e.
M1 + m 1
M2 + m 2
a
a
Proof. From the first inequality in (4.109), we get
kf (t)k2 + m1 M1 (M1 + m1 ) Re hf (t) , ei
implying
(4.113)
M1 + m 1
kf (t)k2 p
+ m 1 M1
Re hf (t) , ei
m 1 M1
m 1 M1
for a.e. t [a, b] .

Since, obviously,
(4.114)
kf (t)k2 p
2 kf (t)k
+ m 1 M1 ,
m 1 M1
hence, by (4.113) and (4.114)
2 m 1 M1
(4.115)
0
M1 + m 1
for a.e. t [a, b] .
Using the same argument as in the proof of Corollary 46, we deduce
the desired inequality. We omit the details.
183
4.5.2. The Case of Orthonormal Vectors. The following result holds [5].
Theorem 65 (Dragomir, 2004). Let {e1 , . . . , en } be a family of
orthonormal vectors in the complex Hilbert space (H; h, i). If kj , hj
0, j {1, . . . , n} and f L ([a, b] ; H) are such that
(4.116)
kj kf (t)k Re hf (t) , ej i ,
hj kf (t)k Im hf (t) , ej i
for each j {1, . . . , n} and a.e. t [a, b] , then

# 12 Z
" n
Z b

b
X

2
2
.
(4.117)
kj + hj
kf (t)k dt
f
(t)
dt

a
j=1

Z b
X
Z b
n
(4.118)
f (t) dt =
kf (t)k dt
(kj + ihj ) ej .
a
j=1
Proof. Before we prove the theorem, let us recall that, if x H

and e1 , . . . , en are orthonormal vectors, then the following identity holds
true:

2
n
n

X
X

2
(4.119)
hx, ej i ej = kxk
|hx, ej i|2 .
x

j=1
j=1
As a consequence of this identity, we have the Bessel inequality

n
X
(4.120)
|hx, ej i|2 kxk2 , x H,
j=1
in which, the case of equality holds if and only if

(4.121)
x=
n
X
hx, ej i ej .
j=1
Now, applying Bessels inequality for x =

sively
Z b
2

(4.122)
f
(t)
dt

Rb
a
f (t) dt, we have succes-
2
2 X
n Z b
n Z b
X

=
hf
(t)
,
e
i
dt
f
(t)
dt,
e
j
j

j=1
j=1
184
4.
Z b
2
n Z b
X

=
Re
hf
(t)
,
e
i
dt
+
i
Im
hf
(t)
,
e
i
dt
j
j

a
a
j=1
" Z
2 Z b
2 #
n
b
X
=
Re hf (t) , ej i dt +
Im hf (t) , ej i dt
.
a
j=1
Integrating (4.116) on [a, b] , we get

Z b
Z b
(4.123)
Re hf (t) , ej i dt kj
kf (t)k dt
a
and
Z
Z
Im hf (t) , ej i dt hj
(4.124)
kf (t)k dt
for each j {1, . . . , n} .

Squaring and adding the above two inequalities (4.123) and (4.124),
we deduce
" Z
2 #
2 Z b
n
b
X
Im hf (t) , ej i dt
Re hf (t) , ej i dt +
j=1
n
X
kj2
h2j
Z
2
kf (t)k dt ,
j=1
which combined with (4.122) will produce the desired inequality (4.117).
Now, if (4.118) holds true, then

Z b
Z b

n
X

kf (t)k dt
(kj + ihj ) ej
f (t) dt =

a
a
j=1
2 21

Z b
n
X

=
kf (t)k dt
(kj + ihj ) ej

a
j=1
Z
"
kf (t)k dt
n
X
# 12

kj2 + h2j
j=1
and the case of equality holds in (4.117).

Conversely, if the equality holds in (4.117), then it must hold in all
the inequalities used to prove (4.117) and therefore we must have
Z b
2 X
2
n Z b

(4.125)
f
(t)
dt
=
f
(t)
dt,
e
j

a
j=1
185
and
(4.126)
kj kf (t)k = Re hf (t) , ej i and hj kf (t)k = Re hf (t) , ej i
for each j {1, . . . , n} and a.e. t [a, b] .

From (4.125), on using the identity (4.121), we deduce that

Z b
n Z b
X
(4.127)
f (t) dt =
f (t) dt, ej ej .
a
j=1
Now, multiplying the second equality in (4.126) with the imaginary

unit i, integrating both inequalities on [a, b] and summing them up, we
get
Z b

Z b
(4.128)
(kj + ihj )
kf (t)k dt =
f (t) dt, ej
a
for each j {1, . . . , n} .

Finally, utilising (4.127) and (4.128), we deduce (4.118) and the
theorem is proved.
The following corollaries are of interest [5].
Corollary 48. Let e1 , . . . , em be orthonormal vectors in the complex Hilbert space (H; h, i) and k , k (0, 1) , k {1, . . . , n} . If
f L ([a, b] ; H) is such that
kf (t) ek k k ,
kf (t) iek k k
for each k {1, . . . , n} and for a.e. t [a, b] , then we have the
inequality
" n
# 21 Z
Z b

b
X

2
2

(4.129)
2 k k
kf (t)k dt
f (t) dt
.
a
k=1

Z b
(4.130)
f (t) dt
a
Z
=
a

X
n q
q
1 2k + i 1 2k ek .
kf (t)k dt
k=1
The proof follows by Theorem 65 and is similar to the one from

Corollary 46. We omit the details.
Next, the following result may be stated [5]:
186
4.
Corollary 49. Let e1 , . . . , em be as in Corollary 48 and Mk

mk > 0, Nk nk > 0, k {1, . . . , n} . If f L ([a, b] ; H) is such that
either
Re hMk ek f (t) , f (t) mk ek i 0,
Re hNk iek f (t) , f (t) nk iek i 0
or, equivalently,

M
+
m
k
k

f (t)
e
k

2

f (t) Nk + nk iek

2
1
(Mk mk ) ,
2
1
(Nk nk )
2
for each k {1, . . . , n} and a.e. t [a, b] , then we have the inequality
(4.131) 2
( m
X
k=1
m k Mk
n k Nk
2 +
(Mk + mk )
(Nk + nk )2
) 12 Z
kf (t)k dt
a
Z b

f (t) dt
.
a

Z b

Z b
(4.132)
f (t) dt = 2
kf (t)k dt
a
n
X
m k Mk
k=1
n k Nk
+i
Mk + m k
Nk + n k

ek .
The proof employs Theorem 65 and is similar to the one in Corollary

47. We omit the details.
4.6. Applications for Complex-Valued Functions
The following proposition holds [2].
Proposition 48. If f : [a, b] C is a Lebesgue integrable function
with the property that there exists a constant K 1 such that
(4.133)
|f (t)| K [ Re f (t) + Im f (t)]
for a.e. t [a, b] , where , R, 2 + 2 = 1 are given, then we have

the following reverse of the continuous triangle inequality:
Z b

Z b

(4.134)
|f (t)| dt K
f (t) dt .
a
4.6. APPLICATIONS FOR COMPLEX-VALUED FUNCTIONS
187

Z b
Z b
1
f (t) dt =
( + i)
|f (t)| dt.
K
a
a
The proof is obvious by Theorem 57, and we omit the details.
Remark 54. If in the above Proposition 48 we choose = 1, = 0,
then the condition (4.133) for Re f (t) > 0 is equivalent to
[Re f (t)]2 + [Im f (t)]2 K 2 [Re f (t)]2
or with the inequality:
|Im f (t)| 2
K 1.
Re f (t)
Now, if we assume that
(4.135)
|arg f (t)| ,

0,
,
2

then, for Re f (t) > 0,

|tan [arg f (t)]| =
and if we choose K =
1
cos
|Im f (t)|
tan ,
Re f (t)
> 1, then
K 2 1 = tan ,
and by Proposition 48, we deduce

Z b

Z b

(4.136)
cos
|f (t)| dt
f (t) dt ,
a
which is exactly the Karamata inequality (4.2) from the Introduction.

Obviously, the result from Proposition 48 is more comprehensive
since for other values of (, ) R2 with 2 + 2 = 1 we can get
different sufficient conditions for the function f such that the inequality
(4.134) holds true.
A different sufficient condition in terms of complex disks is incorporated in the following proposition [2].
Proposition 49. Let e = + i with 2 + 2 = 1, r (0, 1) and
f : [a, b] C a Lebesgue integrable function such that
(e, r) := {z C| |z e| r} for a.e. t [a, b] .
(4.137) f (t) D
Z b

Z b

2

(4.138)
1r
|f (t)| dt
f (t) dt .
a
188
4.

Z b
Z b
2
f (t) dt = 1 r ( + i)
|f (t)| dt.
a
The proof follows by Corollary 33 and we omit the details.

Further, we may state the following proposition as well [2].
Proposition 50. Let e = +i with 2 + 2 = 1 and M m > 0.
If f : [a, b] C is such that
h

i
(4.139)
Re (M e f (t)) f (t) me 0 for a.e. t [a, b] ,
or, equivalently,

1
M
+
m
f (t)
(M m)
(4.140)
e

2
2
for a.e. t [a, b] ,
Z b

Z

2 mM b

(4.141)
|f (t)| dt
f (t) dt ,
M +m a
a
or, equivalently,
(4.142)
Z b

(0 )
|f (t)| dt
f (t) dt
a
a

2 Z

M m b

.
f
(t)
dt

M +m
Z
The equality holds in (4.141) (or in the second part of (4.142)) if and
only if
Z b
Z b
2 mM
f (t) dt =
( + i)
|f (t)| dt.
M +m
a
a
The proof follows by Corollary 34 and we omit the details.
Remark 55. Since
M e f (t) = M Re f (t) + i [M Im f (t)] ,
f (t) me = Re f (t) m i [Im f (t) m]
hence
h

i
(4.143) Re (M e f (t)) f (t) me
= [M Re f (t)] [Re f (t) m]
+ [M Im f (t)] [Im f (t) m] .
189
It is obvious that, if
(4.144)
m Re f (t) M
for a.e. t [a, b] ,
m Im f (t) M
for a.e. t [a, b] ,
and
(4.145)
then, by (4.143),
i
h

Re (M e f (t)) f (t) me 0
for a.e. t [a, b] ,
and then either (4.141) or (4.144) hold true.

We observe that the conditions (4.144) and (4.145) are very easy
to verify in practice and may be useful in various applications where
reverses of the continuous triangle inequality are required.
Remark 56. Similar results may be stated for functions f : [a, b]
R or f : [a, b] H, with H particular instances of Hilbert spaces of
significance in applications, but we leave them to the interested reader.
n
Let e = + i (, R) be a complex number with the property

that |e| = 1, i.e., 2 + 2 = 1. The following proposition concerning a reverse of the continuous triangle inequality for complex-valued
functions may be stated [3]:
Proposition 51. Let f : [a, b] C be a Lebesgue integrable function with the property that there exists a constant (0, 1) such that
(4.146)
|f (t) e| for a.e. t [a, b] ,
where e has been defined above. Then we have the following reverse of
the continuous triangle inequality
Z b

Z b

(4.147)
|f (t)| dt
f (t) dt
(0 )
a

p
1 2 1 + 1 2
Z b

Z b

Re f (t) dt +
Im f (t) dt .
a
The proof follows by Corollary 37, and the details are omitted.
On the other hand, the following result is perhaps more useful for
applications [3]:
190
4.
Proposition 52. Assume that f and e are as in Proposition 51.

If there exists the constants M m > 0 such that either
i
h

(4.148)
Re (M e f (t)) f (t) me 0
or, equivalently,

f (t) M + m e 1 (M m)
2

2
(4.149)
for a.e. t [a, b] , holds, then

Z b

Z b

(4.150) (0 )
|f (t)| dt
f (t) dt
a
2

Z b
M m Z b
Re f (t) dt +
Im f (t) dt .
2 Mm
a
a
The proof may be done on utilising Corollary 38, but we omit the
details
Subsequently, on making use of Corollary 40, one may state the
following result as well [3]:
Proposition 53. Let f be as in Proposition 51 and the measurable
functions K, k : [a, b] [0, ) with the property that
(K k)2
L [a, b]
K +k
and
k (t) Re f (t) K (t) and k (t) Im f (t) K (t)
for a.e. t [a, b] , where , are assumed to be positive and satisfying
the condition 2 + 2 = 1. Then the following reverse of the continuous
triangle inequality is valid:
Z b

Z b

|f (t)| dt
f (t) dt
(0 )
a
a
Z b
1
[K (t) k (t)]2
dt.
4 a K (t) + k (t)
by a smaller quantity.
Remark 57. One may realise that similar results can be stated if
the Corollaries 41-44 obtained above are used. For the sake of brevity,
we do not mention them here.
191
Let f : [a, b] C be a Lebesgue integrable function and M

1 m 0. The condition (4.76) from Theorem 61, which plays a
fundamental role in the results obtained above, can be translated in
this case as
h

i
(4.151)
Re (M f (s) f (t)) f (t) mf (s) 0
for a.e. a t s b.
Since, obviously
h

i
Re (M f (s) f (t)) f (t) mf (s)
= [(M Re f (s) Re f (t)) (Re f (t) m Re f (s))]
+ [(M Im f (s) Im f (t)) (Im f (t) m Im f (s))]
hence a sufficient condition for the inequality in (4.151) to hold is
(4.152)
m Re f (s) Re f (t) M Re f (s)
and
m Im f (s) Im f (t) M Im f (s)
for a.e. a t s b.
Utilising Theorems 61, 62 and 63 we may state the following results
incorporating quadratic reverses of the continuous triangle inequality
[4]:
Proposition 54. With the above assumptions for f, M and m, and
if (4.151) holds true, then we have the inequalities
2
Z b
2 Z b

|f (t)| dt
f (t) dt
a
a
Z
1 (M m)2 b
+
(s a) |f (s)|2 ds,
2
M +m a

1 Z
Z b

M + m 2 b
|f (t)| dt
f (t) dt ,

2 Mm
a
a
and
Z
2
Z b

M + m b
2
.
[(b s) + mM (s a)] |f (s)| ds
f
(s)
ds

2
a
Remark 58. One may wonder if there are functions satisfying the
condition (4.152) above. It suffices to find examples of real functions
: [a, b] R verifying the following double inequality
(4.153)
(s) (t) (s)
192
4.
for some given , with 0 1 < for a.e. a t s b.

For this purpose, consider : [a, b] R a differentiable function
on (a, b), continuous on [a, b] and with the property that there exists
0 such that
(4.154)
0 (u) for any u (a, b) .
By Lagranges mean value theorem, we have, for any a t s b

(s) (t) = 0 () (s t)
with t s. Therefore, for a t s b, by (4.154), we have the
inequality
(b a) (s t) (s) (t) (s t) (b a) .
If we choose the function : [a, b] R given by
(t) := exp [ (t)] , t [a, b] ,
and := exp [ (b a)] 1, := exp [ (b a)] , then (4.153) holds
true for any a t s b.
The following reverse of the continuous triangle inequality for complexvalued functions that improves Karamatas result (4.1) holds [5].
Proposition 55. Let f L ([a, b] ; C) with the property that
(4.155)
0 1 arg f (t) 2 <
2
for a.e. t [a, b] . Then we have the inequality
Z b

Z b
q

2
2
(4.156)
sin 1 + cos 2
|f (t)| dt
f (t) dt .
a

Z b
Z b
(4.157)
f (t) dt = (cos 2 + i sin 1 )
|f (t)| dt.
a
Proof. Let f (t) = Re f (t) + i Im f (t) . We may assume that

Re f (t) 0, Im f (t) > 0, for
a.e. t [a, b] , since, by (4.155),
Im f (t)
= tan [arg f (t)] 0, 2 , for a.e. t [a, b] . By (4.155), we

Re f (t)
obviously have

2
Im f (t)
2
0 tan 1
tan2 2 ,
Re f (t)
for a.e. t [a, b] , from where we get
1
[Im f (t)]2 + [Re f (t)]2
,
2
cos 2 2
[Re f (t)]
193
for a.e. t [a, b] , and

1
[Im f (t)]2 + [Re f (t)]2
1 + tan2 1
=
,
2
2
tan 1
sin 1
[Im f (t)]
for a.e. t [a, b] , giving the simpler inequalities
|f (t)| cos 2 Re (f (t)) ,
|f (t)| sin 1 Im (f (t))
for a.e. t [a, b] .

Now, applying Theorem 64 for the complex Hilbert space C endowed with the inner product hz, wi = z w for k1 = cos 2 , k2 = sin 1
and e = 1, we deduce the desired inequality (4.156). The case of equality is also obvious and we omit the details.
Another result that has an obvious geometrical interpretation is the
following one [5].
Proposition 56. Let e C with |e| = 1 and 1 , 2 (0, 1) . If
f (t) L ([a, b] ; C) such that
(4.158)
|f (t) e| 1 ,
|f (t) ie| 2
for a.e. t [a, b] ,

Z b

Z b
q

2
2
(4.159)
2 1 2
|f (t)| dt
f (t) dt ,
a

q
Z b
Z b
q
(4.160)
f (t) dt =
1 21 + i 1 22
|f (t)| dt e.
a
The proof is obvious by Corollary 46 applied for H = C and we

omit the details.
Remark 59. If we choose e = 1, and for 1 , 2 (0, 1) we define
(1, 1 ) := {z C| |z 1| 1 } , D
(i, 2 ) := {z C| |z i| 2 } ,
D
then obviously the intersection domain
(1, 1 ) D
(i, 2 )
S1 ,2 := D
is nonempty if and only if 1 + 2 > 2.

If f (t) S1 ,2 for a.e. t [a, b] , then (4.159) holds true. The
equality holds in (4.159) if and only if
q
Z b
Z b
q
2
2
f (t) dt =
1 1 + i 1 2
|f (t)| dt.
a
Bibliography
Hilbert and Banach spaces, Proceedings Amer. Math. Soc., 17(1) (1966), 88-97.
[2] S.S. DRAGOMIR, Reverses of the continuous triangle inequality for
Bochner integral of vector valued function in Hilbert spaces. Preprint,
RGMIA Res. Rep. Coll., 7(2004), Supplement, Article 11, [Online
[3] S.S. DRAGOMIR, Additive reverses of the continuous triangle inequality for Bochner integral of vector valued functions in Hilbert spaces.
Preprint, RGMIA Res. Rep. Coll., 7(2004), Supplement, Article 12, [Online
[4] S.S. DRAGOMIR, Quadratic reverses of the continuous triangle inequality for Bochner integral of vector-valued functions in Hilbert spaces,
[5] S.S. DRAGOMIR, Some reverses of the continuous triangle inequality for
Bochner integral of vector-valued functions in complex Hilbert spaces,
[6] J. KARAMATA, Teorija i Praksa Stieltjesova Integrala (Serbo-Croatian)
(Stieltjes Integral, Theory and Practice), SANU, Posebna izdanja, 154,
Beograd, 1949.
[7] M. MARDEN, The Geometry of the Zeros of a Polynomial in a Complex Variable, Amer. Math. Soc. Math. Surveys, 3, New York, 1949.
J.E. PECARI
and A.M. FINK, Classical and

C
New Inequalities in Analysis, Kluwer Academic Publishers, Dordrecht/Boston/London, 1993.
[9] M. PETROVICH, Module dune somme, L Ensignement Mathematique, 19
(1917), 53-56.
[10] H.S. WILF, Some applications of the inequality of arithmetic and geometric
means to polynomial equations, Proceedings Amer. Math. Soc., 14 (1963),
263-265.
195
CHAPTER 5
Reverses of the CBS and Heisenberg Inequalities

5.1. Introduction
Assume that (K; h, i) is a Hilbert space over the real or complex number field K. If : [a, b] R [0, ) is a Lebesgue inteRb
grable function with a (t) dt = 1, then we may consider the space
L2 ([a, b] ; K) of all functionsf : [a, b] K, that are Bochner measurRb
able and a (t) kf (t)k2 dt < . It is well known that L2 ([a, b] ; K)
endowed with the inner product h, i defined by
b
Z
hf, gi :=
(t) hf (t) , g (t)i dt

a

Z
kf k :=
12
(t) kf (t)k dt ,
2

The following integral inequality is known in the literature as the
Cauchy-Bunyakovsky-Schwarz (CBS) inequality
Z
b
2
(t) kf (t)k dt
(5.1)
a
(t) kg (t)k2 dt
Z b
2

(t) hf (t) , g (t)i dt ,
a
provided f, g L2 ([a, b] ; K) .
The case of equality holds in (5.1) iff there exists a constant K
such that f (t) = g (t) for a.e. t [a, b] .
197
198
5. CBS AND HEISENBERG INEQUALITIES
Another version of the (CBS) inequality for one vector-valued and

one scalar function is incorporated in:
Z b
Z b
2
(5.2)
(t) | (t)| dt
(t) kf (t)k2 dt
a
Z b
2

(t) (t) f (t) dt
,
a
L2
L2
provided
([a, b]) and f
([a, b] ; K) , where L2 ([a, b]) denotes
Rb
the Hilbert space of scalar functions for which a (t) | (t)|2 dt < .
The equality holds in (5.2) iff there exists a vector e K such that
f (t) = (t)e for a.e. t [a, b] .
In this chapter some reverses of the inequalities (5.1) and (5.2) are
given under various assumptions for the functions involved. Natural
applications for the Heisenberg inequality for vector-valued functions
in Hilbert spaces are also provided.
5.2. Some Reverse Inequalities
5.2.1. The General Case. The following result holds [1].
Theorem 66 (Dragomir, 2004). Let f, g L2 ([a, b] ; K) and r > 0
be such that
kf (t) g (t)k r kg (t)k
(5.3)
for a.e. t [a, b] . Then we have the inequalities:

Z b
Z b
2
(5.4)
0
(t) kf (t)k dt
(t) kg (t)k2 dt
a
Z b
2

a
Z b
Z b
2
(t) kf (t)k dt
(t) kg (t)k2 dt
Z
2
(t) Re hf (t) , g (t)i dt
r2
(t) kf (t)k2 dt.
The constant C = 1 in front of r2 is best possible in the sense that it

5.2. SOME REVERSE INEQUALITIES
199
Proof. We will use the following result obtained in [2]:

In the inner product space (H; h, i) , if x, y H and r > 0 are
such that kx yk r kyk , then
0 kxk2 kyk2 |hx, yi|2
(5.5)
kxk2 kyk2 [Re hx, yi]2 r2 kxk2 .

The constant c = 1 in front of r2 is best possible in the sense that it
If (5.3) holds, true, then
kf
gk2
(t) kf (t) g (t)k dt r
(t) dt = r2
and
kgk2
b
2
(t) kg (t)k dt r
(t) dt = r2
and thus kf gk r kgk . Applying the inequality (5.5) for

L2 ([a, b] ; K) , h, ip , we deduce the desired inequality (5.4).
1
If we choose (t) = ba
, f (t) = x, g (t) = y, x, y K, t [a, b] ,
then from (5.4) we recapture (5.5) for which the constant c = 1 in front
of r2 is best possible.
We next point out some general reverse inequalities for the second
(CBS) inequality (5.2) [1].
Theorem 67 (Dragomir, 2004). Let L2 ([a, b]) , g L2 ([a, b] ; K)
and a K, r > 0 such that kak > r. If the following condition holds
(5.6)
kg (t)
(t) ak r | (t)|
for a.e. t [a, b] , (note that, if (t) 6= 0 for a.e. t [a, b] , then the
condition (5.6) is equivalent to
(5.7)

g (t)

a

r
(t)
200
for a.e. t [a, b]), then we have the following inequality

Z
(5.8)
12
(t) kg (t)k dt
2
(t) | (t)| dt
a
Z b

1
q
Re
(t) (t) g (t) dt, a
2
a
2
kak r
Z b

kak

q
(t) (t) g (t) dt

;
2
a
2
kak r
a
12
(5.9) 0
(t) | (t)| dt
(t) kg (t)k dt
a
a
Z b

(t) (t) g (t) dt

a
Z b
12
Z b
(t) | (t)|2 dt
(t) kg (t)k2 dt
a
a
Z b

a
Re
(t) (t) g (t) dt,
kak
a
2
r

q
q
2
2
2
2
kak r kak + kak r
Z b

a
Re
(t) (t) g (t) dt,
kak
a
Z b

2

r

q
(t) (t) g (t) dt
q

;
2
2
a
2
2
kak r kak + kak r
Z
(t) | (t)| dt
(5.10)
a
(t) kg (t)k2 dt
1
kak r2
2
Z b
2
Re
(t) (t) g (t) dt, a
a
Z b
2

kak

,
(t)
(t)
g
(t)
dt

kak2 r2 a
2
201
and
Z
0
(5.11)
b
2
(t) kg (t)k2 dt
(t) | (t)| dt
a
Z b
2

(t)
(t)
g
(t)
dt

a
Z b
Z b
(t) | (t)|2 dt
(t) kg (t)k2 dt
a
Z b
2
a
Re
(t) (t) g (t) dt,
kak
a

2
Z
b
r2
Re
(t) (t) g (t) dt, a

kak2 kak2 r2
a
Z b
2

r2

(t) (t) g (t) dt

2

.
kak r2 a
All the inequalities (5.8) (5.11) are sharp.
Proof. From (5.6) we deduce
kg (t)k2 2 Re hg (t) ,
(t) ai + | (t)|2 kak2 | (t)|2 r2
for a.e. t [a, b] , which is clearly equivalent to:

kg (t)k2 + kak2 r2 | (t)|2 2 Re h (t) g (t) , ai
(5.12)
for a.e. t [a, b] .

If we multiply (5.12) by (t) 0 and integrate over t [a, b] , then
we deduce
Z
b
2
(t) | (t)|2 dt
a

Z b
(t) (t) g (t) dt, a .
2 Re
(t) kg (t)k dt + kak r
(5.13)
202
kak2 r2 > 0, we get
Now, dividing (5.13) by

(5.14) q
1
kak2 r2
(t) kg (t)k2 dt
Z b
q
2
+ kak r2
(t) | (t)|2 dt
a
Z b

2
q
Re
(t) (t) g (t) dt, a .
2
a
2
kak r
1
p + q 2 pq,
> 0, p, q 0,
we can state that

s
s
Z b
Z b
2
(t) | (t)| dt
(t) kg (t)k2 dt
(5.15) 2
a
kak2 r2
(t) kg (t)k2 dt
kak
Z
r2
(t) | (t)|2 dt.
Making use of (5.14) and (5.15), we deduce the first part of (5.8).
The second part of (5.8) is obvious by Schwarzs inequality

Z b
Z b

Re
(t) (t) g (t) dt, a
(t) (t) g (t) dt
kak .
a
If (t) =
1
,
ba
(t) = 1, g (t) = x K, then, from (5.8) we get

kxk kak
,
Re hx, ai q
kak2 r2
kak2 r2
kxk q
provided kx ak r < kak , x, a K. The sharpness of this inequality

has been shown in [2], and we omit the details.
The other inequalities are obvious consequences of (5.8) and we
omit the details.
203
5.2.2. Some Particular Cases. It has been shown in [2] that, for
A, a K (K = C, R) and x, y H, where (H; h, i) is an inner product
over the real or complex number field K, the following inequality holds

hx, yi
1 Re A + a
kxk kyk
(5.16)
1
2
[Re (A
a)] 2
1
|A + a|
1 |hx, yi|
2 [Re (A
a)] 2
provided Re (A
a) > 0 and
Re hAy x, x ayi 0,
(5.17)
or, equivalently,
(5.18)

x a + A y 1 |A a| kyk ,

2
2
holds. The constant 12 is best possible in (5.16).

From (5.16), we can deduce the following results
(5.19)
0 kxk kyk Re hx, yi

h

i
1
+a
2
Re
A
2
[Re
(A
a
)]
hx,
yi
1

1
2
[Re (A
a)] 2

1

2
2 [Re (A
a)]
1 A + a

|hx, yi|
1
2
[Re (A
a)] 2
and
(5.20)
0 kxk kyk |hx, yi|

1
1 |A + a| 2 [Re (A
a)] 2

|hx, yi| .
1
2
[Re (A
a)] 2
If one assumes that A = M, a = m, M m > 0, then, from (5.16),
(5.19) and (5.20) we deduce the much simpler and more useful results:
(5.21)
kxk kyk
1 M +m
Re hx, yi ,

2
Mm

(5.22)
0 kxk kyk Re hx, yi
2
M m
Re hx, yi
Mm
204
and

(5.23)
0 kxk kyk |hx, yi|
2
M m
|hx, yi| ,
Mm
provided
Re hM y x, x myi 0
or, equivalently
(5.24)

1

M
+
m
(M m) kyk .
x
y
2

2
Squaring the second inequality in (5.16), we can get the following results as well:
(5.25)
1 |A a|2
0 kxk kyk |hx, yi|
|hx, yi|2 ,
4 Re (A
a)
2
provided (5.17) or (5.16) holds. Here the constant 14 is also best possible.
Using the above inequalities for vectors in inner product spaces,
we are able to state the following theorem concerning reverses of the
(CBS) integral inequality for vector-valued functions in Hilbert spaces
[1].
Theorem 68 (Dragomir, 2004). Let f, g L2 ([a, b] ; K) and ,
K with Re (
) > 0. If
(5.26)
Re hg (t) f (t) , f (t) g (t)i 0
for a.e. t [a, b] , or, equivalently,

1
f (t)
(5.27)
g (t)

2 | | kg (t)k
2
for a.e. t [a, b] , then we have the inequalities
Z b
12
Z b
2
2
(5.28)
(t) kf (t)k dt
(t) kg (t)k dt
a
a
h
i
Rb
1 Re + a (t) hf (t) , g (t)i dt

1
2
[Re (
)] 2
Z

1
| + | b
(t) hf (t) , g (t)i dt ,

1
2 [Re (
)] 2 a
(5.29)
12 Z b
12
(t) kf (t)k2 dt
(t) kg (t)k2 dt
Z
205
0
a

a
hn
oR
i
1
b
+ 2 [Re (
2
Re
)]
(t)
hf
(t)
,
g
(t)i
dt
a
1

1
2
[Re (
)] 2

1

2 Z b
2
[Re
(
)]

1

,
(t)
hf
(t)
,
g
(t)i
dt

1

2
a
[Re (
)] 2
b
Z
(5.30)
21 Z b
12
2
(t) kf (t)k dt
(t) kg (t)k dt
a
Z b

2
0
a

1 Z

)] 2 b
1 | + | 2 [Re (
,

(t)
hf
(t)
,
g
(t)i
dt
1

2
2
a
[Re (
)]
and
Z
(5.31)
b
2
(t) kf (t)k dt
a
(t) kg (t)k2 dt
Z b
2

a
2
2 Z b

1 | |

(t) hf (t) , g (t)i dt .

4 Re (
) a
The constants
1
2
and
1
4
above are sharp.
In the case where , are positive real numbers, the following corollary incorporating more convenient reverses for the (CBS) integral inequality, may be stated [1].
Corollary 50. Let f, g L2 ([a, b] ; K) and M m > 0. If
(5.32)
Re hM g (t) f (t) , f (t) mg (t)i 0

1

m+M

g (t)
(5.33)
f (t)
2 (M m) kg (t)k
2
206

Z
(5.34)
(t) kf (t)k2 dt
12
(t) kg (t)k2 dt
1 M +m

2
mM
b
Z
(5.35)
(t) Re hf (t) , g (t)i dt,

a
21 Z b
12
2
(t) kf (t)k dt
(t) kg (t)k dt
2
0
a
a
b

2 Z
b
M
m
1

(t) Re hf (t) , g (t)i dt,
2
mM
a
a
Z
(5.36)
21 Z b
12
2
(t) kf (t)k dt
(t) kg (t)k dt
a
Z b

2
0
a

1

2
2
M m
mM
Z b

(t) hf (t) , g (t)i dt ,

a
and
Z
(5.37)
(t) kf (t)k dt
a
(t) kg (t)k2 dt
Z b
2

a
Z
2

1 (M m)2 b
.

(t)
hf
(t)
,
g
(t)i
dt

4
mM
a
The constants
1
2
and
1
4
above are best possible.
On utilising the general result of Theorem 67, we are able to state a

number of interesting reverses for the (CBS) inequality in the case when
one function takes vector-values while the other is a scalar function [1].
207
Theorem 69 (Dragomir, 2004). Let L2 ([a, b]) , g L2 ([a, b] ; K) ,

e K, kek = 1, , K with Re (
) > 0. If

1

| | | (t)|
g (t)
(5.38)
(t)
e

2 2
for a.e. t [a, b] , or, equivalently
(5.39)
Re h
(t) e g (t) , g (t)
(t) ei 0
for a.e. t [a, b] , (note that, if (t) 6= 0 for a.e. t [a, b] , then (5.38)
is equivalent to

g (t) + 1

(5.40)
e | |

(t)
2 2
for a.e. t [a, b] , and (5.39) is equivalent to
*
+
g (t) g (t)
(5.41)
Re e
,
e 0
(t) (t)
for a.e. t [a, b]), then the following reverse inequalities are valid:
12
(t) | (t)| dt
(t) kg (t)k dt
a
a
h
Ei
DR b
+
Re
(t)
(t)
g
(t)
dt,
e
a
Z
(5.42)
2 [Re (
)] 2
Z b

1
| + |

(t) (t) g (t) dt

1
;
2 [Re (
a
)] 2
Z
(5.43)
12
(t) | (t)| dt
(t) kg (t)k dt
a
Z b

(t) (t) g (t) dt

2
0
a
Z
21
(t) | (t)| dt
(t) kg (t)k dt
a

Z b

+
Re
(t) (t) g (t) dt, e
| + |
a
2
208
| |2

p
p
2 Re (
) | + | + 2 Re (
)

Z b

+
Re
(t) (t) g (t) dt, e
| + |
a
Z b

| |2
;

p
(t)
(t)
g
(t)
dt
p

a
2 Re (
) | + | + 2 Re (
)
b
(t) | (t)| dt
(5.44)
a
(t) kg (t)k2 dt
1
1
4 Re (
)

Z b
2

Re +
(t) (t) g (t) dt, e
a
Z b
2

1 | + |

(t) (t) g (t) dt

4 Re (
) a
2
and
Z
(5.45)
b
2
(t) | (t)| dt
a
(t) kg (t)k2 dt
Z b
2

(t) (t) g (t) dt

a
Z b
Z b
2
(t) | (t)| dt
(t) kg (t)k2 dt
a

Z b
2
+
Re
(t) (t) g (t) dt, e
| + |
a
1
| |2
4 | + |2 Re (
)

Z b
2

Re +
(t) (t) g (t) dt, e
a
Z b
2

1 | |

.

(t)
(t)
g
(t)
dt

4 Re (
) a
2
The constants
1
2
and
1
4
above are sharp.
In the particular case of positive constants, the following simpler

version of the above inequalities may be stated.
209
Corollary 51. Let L2 ([a, b]) \ {0} , g L2 ([a, b] ; K) , e K,

kek = 1 and M, m R with M m > 0. If

g (t) M + m 1

e

2 (M m)
(t)
2
(5.46)

g (t) g (t)
Re M e
,
me
(t)
(t)
(5.47)

0
for a.e. t [a, b] , then we have

Z
b
2
(5.48)
a
1
2
1
2
12
(t) kg (t)k dt
(t) | (t)| dt
a
Z b

M +m

Re
(t) (t) g (t) dt, e
Mm
a
Z b

M +m

;

(t)
(t)
g
(t)
dt

Mm a
Z
(5.49)
12
(t) | (t)| dt
(t) kg (t)k dt
a
Z b

(t) (t) g (t) dt

Z
0
a
Z
12
(t) | (t)| dt
(t) kg (t)k dt
a
Z b

Re
(t) (t) g (t) dt, e
Z
2
M m
2 Mm

2
M m
2 Mm
Z
Re

(t) (t) g (t) dt, e
Z b

(t) (t) g (t) dt

a
210
Z
(5.50)
(t) kg (t)k2 dt
(t) | (t)| dt
a
Z b
2
(M + m)
Re
(t) (t) g (t) dt, e
Mm
a
Z b
2

(M + m)2

(t) (t) g (t) dt

Mm
a
2
and
Z
(5.51)
(t) | (t)| dt
a
(t) kg (t)k2 dt
Z b
2

(t) (t) g (t) dt

a
Z b
Z b
2
(t) | (t)| dt
(t) kg (t)k2 dt
a
Z b
2
Re
(t) (t) g (t) dt, e
a
The constants
1
2
and
Z b
2
(M m)
Re
(t) (t) g (t) dt, e
Mm
a
Z b
2

(M m)2

.
(t)
(t)
g
(t)
dt

Mm a
2
1
4
above are sharp.
5.2.3. Reverses of the Heisenberg Inequality. It is well known

that if (H; h, i) is a real or complex Hilbert space and f : [a, b]
R H is an absolutely continuous vector-valued function, then f is
differentiable almost everywhere on [a, b] , the derivative f 0 : [a, b] H
is Bochner integrable on [a, b] and
Z t
(5.52)
f (t) =
f 0 (s) ds
for any t [a, b] .
a
The following theorem provides a version of the Heisenberg inequalities in the general setting of Hilbert spaces [1].
Theorem 70 (Dragomir, 2004). Let : [a, b] H be an absolutely
continuous function with the property that b k (b)k2 = a k (a)k2 .
Then we have the inequality:
Z b
2
Z b
Z b
2
2
2
2
(5.53)
k (t)k dt 4
t k (t)k dt
k0 (t)k dt.
a
211

Proof. Integrating by parts, we have successively
Z
k (t)k2 dt
a
b Z b

d
2
k (t)k2 dt
= t k (t)k
t
dt
a
a
Z b
d
2
2
= b k (b)k a k (a)k
t h (t) , (t)i dt
dt
a
Z b
=
t [h0 (t) , (t)i + h (t) , 0 (t)i] dt
a
Z b
= 2
t Re h0 (t) , (t)i dt
a
Z b
Re h0 (t) , (t) (t)i dt.
=2
(5.54)
If we apply the (CBS) integral inequality

Z
Z
Re hg (t) , h (t)i dt
b
2
kg (t)k dt
a
12
kh (t)k dt
2
for g (t) = 0 (t) , h (t) = t (t) , t [a, b] , then we deduce the desired
inequality (5.53).
The fact that 4 is the best possible constant in (5.53) follows from
the fact that in the (CBS) inequality, the case of equality holds iff
g (t) = h (t) for a.e. t [a, b] and a given scalar in K. We omit the
details.
For details on the classical Heisenberg inequality, see, for instance,
[4].
The following reverse of the Heisenberg type inequality (5.53) holds
[1].
Theorem 71 (Dragomir, 2004). Assume that : [a, b] H is as
in the hypothesis of Theorem 70. In addition, if there exists a r > 0
such that
(5.55)
k0 (t) t (t)k r k0 (t)k
212

Z b
2
Z b
Z b
1
2
2
2
2
0
(5.56) 0
t k (t)k dt
k (t)k dt
k (t)k dt
4
a
a
a
Z b
2
r
t2 k (t)k2 dt.
a
Proof. We observe, by the identity (5.54), that

Z b
2 Z b
2
1
2
0
(5.57)
k (t)k dt =
Re h (t) , t (t)i dt .
4
a
a
Now, if we apply Theorem 66 for the choices f (t) = t (t) , g (t) =
1
0 (t) , and (t) = ba
, then by (5.4) and (5.57) we deduce the desired
inequality (5.56).
Remark 60. Interchanging the place of t (t) with 0 (t) in Theorem 71, we also have
Z b
2
Z b
Z b
1
2
2
2
2
0
(5.58) 0
k (t)k dt
t k (t)k dt
k (t)k dt
4
a
a
a
Z b
2
2
k0 (t)k dt,
a
provided
k0 (t) t (t)k |t| k (t)k
for a.e. t [a, b] , where > 0 is a given positive number.
The following result also holds [1].
Theorem 72 (Dragomir, 2004). Assume that : [a, b] H is as in
the hypothesis of Theorem 70. In addition, if there exists M m > 0
such that
Re hM t (t) 0 (t) , 0 (t) mt (t)i 0
(5.59)

0
1
M
+
m
(t)
(M m) |t| k (t)k
(5.60)
t
(t)

2
2
Z b
Z b
2
2
2
(5.61)
t k (t)k dt
k0 (t)k dt
a
1 (M + m)2
16
Mm
Z
a
2
k (t)k dt
2
5.3. OTHER REVERSES
213
and
Z
t k (t)k dt
(5.62)
a
1
k (t)k dt
4
2
Z
2
k (t)k dt
2
1 (M m)2
16
Mm
Z
2
k (t)k dt
2
respectively.
Proof. We use Corollary 50 for the choices f (t) = 0 (t) , g (t) =
1
t (t) , (t) = ba
, to get
Z
k (t)k dt
a
t2 k (t)k2 dt
(M + m)2
4M m
Z
2
0
Since, by (5.57)
Z
a
Z b
2
2
1
2
Re h (t) , t (t)i dt =
k (t)k dt ,
4
a
0
hence we deduce the desired result (5.61).

The inequality (5.62) follows from (5.61), and we omit the details.
Remark 61. If one is interested in reverses for the Heisenberg
inequality for scalar valued functions, then all the other inequalities
obtained above for one scalar function may be applied as well. For the
sake of brevity, we do not list them here.
5.3. Other Reverses

5.3.1. The General Case. The following result holds [3].
Theorem 73 (Dragomir, 2004). Let f, g L2 ([a, b] ; K) and r > 0
be such that
(5.63)
kf (t) g (t)k r
214
for a.e. t [a, b] . Then we have the inequalities:

Z b
12
Z b
2
2
(5.64)
0
(t) kf (t)k dt
(t) kg (t)k dt
a
a
Z b

a
Z
21
(t) kf (t)k dt
(t) kg (t)k dt
a
Z b

2
Z
b
2
(t) kf (t)k dt
a
12
(t) kg (t)k dt
2

a
1
r2 .
2
1
The constant 2 in front of r2 is best possible in the sense that it cannot
be replaced by a smaller quantity.
Proof. We will use the following result obtained in [2]:
In the inner product space (H; h, i) , if x, y H and r > 0 are
such that kx yk r, then
0 kxk kyk |hx, yi| kxk kyk |Re hx, yi|
1
kxk kyk Re hx, yi r2 .
2
1
2
The constant 2 in front of r is best possible in the sense that it cannot
be replaced by a smaller constant.
If (5.63) holds true, then
Z b
Z b
2
2
2
kf gk =
(t) kf (t) g (t)k dt r
(t) dt = r2
(5.65)
and thus kf gk r.

Applying the inequality (5.65) for L2 ([a, b] ; K) , h, ip , we deduce the desired inequality (5.64).
1
If we choose (t) = ba
, f (t) = x, g (t) = y, x, y K, t [a, b] ,
then from (5.64) we recapture (5.65) for which the constant 12 in front
of r2 is best possible.
5.3. OTHER REVERSES
215
We next point out some general reverse inequalities for the second
CBS inequality (5.2)[3].
Theorem 74 (Dragomir, 2004). Let L2 ([a, b]) , g L2 ([a, b] ; K)
and v K, r > 0. If

g (t)

(5.66)
v r

(t)

Z b
21
Z b
2
2
(5.67)
0
(t) | (t)| dt
(t) kg (t)k dt
a
a
Z b

(t) (t) g (t) dt

a
Z
12
(t) kg (t)k dt
2
(t) | (t)| dt
a
Z b

v

(t) (t) g (t) dt,
kvk
a
Z b
12
Z b
(t) | (t)|2 dt
(t) kg (t)k2 dt
a
a
Z b

v

Re
(t) (t) g (t) dt,
kvk
a
Z b
12
Z b
2
2
(t) | (t)| dt
(t) kg (t)k dt
a
a
Z b

v
Re
(t) (t) g (t) dt,
kvk
a
2 Z b
1 r

(t) | (t)|2 dt.
2 kvk a
a

kg (t)k2 2 Re h (t) g (t) , vi + | (t)|2 kvk2 r2 | (t)|2
(5.68)
kg (t)k2 + | (t)|2 kvk2 2 Re h (t) g (t) , vi + r2 | (t)|2 .
216
If we multiply (5.68) by (t) 0 and integrate over t [a, b] , then

we deduce
Z b
Z b
2
2
(5.69)
(t) kg (t)k dt + kvk
(t) | (t)|2 dt
a
a
Z b

Z b
2
2 Re
(t) (t) g (t) dt, v + r
(t) | (t)|2 dt.
a
Since, obviously
Z
(5.70) 2 kvk
b
2
(t) | (t)| dt
a
12
(t) kg (t)k dt
2
a
b
(t) kg (t)k dt + kvk
(t) | (t)|2 dt,
hence, by (5.69) and (5.70), we deduce

Z
2 kvk
a
21
(t) | (t)| dt
(t) kg (t)k dt
a
Z b

Z b
2
2 Re
(t) (t) g (t) dt, v + r
(t) | (t)|2 dt,
2
which is clearly equivalent with the last inequality in (5.67).

The other inequalities are obvious and we omit the details.
1
Now, if (t) = ba
, (t) = 1, g (t) = x, x K, then, by the last
inequality in (5.67) we get
1
kxk kvk Re hx, vi r2 ,
2
provided kx vk r, for which we know that (see [2]), the constant
is best possible.
1
2
5.3.2. Some Particular Cases of Interest. It has been shown

in [2] that, for , K (K = C or K = R) with 6= and x, y H,
(H; h, i) is an inner product over the real or complex number field K,
such that either
(5.71)
Re hy x, x yi 0,
or, equivalently,
(5.72)

x + y 1 | | kyk ,

2
2
5.3. OTHER REVERSES
217
holds, then one has the following reverse of Schwarzs inequality

0 kxk kyk |hx, yi|

+

kxk kyk Re
hx, yi
| + |

+
kxk kyk Re
hx, yi
| + |
(5.73)
1 | |2
kyk2 .

4 | + |
The constant 14 is best possible in (5.73) in the sense that it cannot be
replaced by a smaller constant.
If we assume that = M, = m with M m > 0, then from
(5.73) we deduce the much simpler and more useful result
(5.74)
0 kxk kyk |hx, yi| kxk kyk |Re hx, yi|

kxk kyk Re hx, yi
1 (M m)2
kyk2 ,
4
Mm
provided (5.71) or (5.72) holds true with M and m instead of and .

Using the above inequalities for vectors in inner product spaces, we
are able to state the following theorem concerning reverses of the CBS
integral inequality for vector-valued functions in Hilbert spaces [3].
Theorem 75 (Dragomir, 2004). Let f, g L2 ([a, b] ; K) and ,
K with 6= . If
Re hg (t) f (t) , f (t) g (t)i 0
(5.75)

1
+

(5.76)
f (t) 2 g (t) 2 | | kg (t)k
Z
(5.77)
0
a
21
(t) kf (t)k dt
(t) kg (t)k dt
a
Z b

2
218
Z
12
(t) kg (t)k2 dt
(t) kf (t)k2 dt
a
a

Z b

+

Re
| + | a
Z b
21
Z b
2
2
(t) kf (t)k dt
(t) kg (t)k dt
a
a

Z b
+
Re
| + | a
Z
1 | |2 b

(t) kg (t)k2 dt.
4 | + | a
The constant
1
4
Proof. Since, by (5.75),

Re hg f, f gi
Z b
=
(t) Re hg (t) f (t) , f (t) g (t)i dt 0,
a
hence, by (5.73) applied for the Hilbert space

we deduce the desired inequality (5.77).
The best constant follows by the fact that
(5.77) and we omit the details.

1
4
L2
([a, b] ; K) ; h, i ,
is a best constant in
Corollary 52. Let f, g L2 ([a, b] ; K) and M m > 0. If

Re hM g (t) f (t) , f (t) mg (t)i 0
(5.78)

1
m+M

(M m) kg (t)k
(5.79)
f
(t)
g
(t)

2
2
for a.e. t [a, b] , then
Z
(5.80)
0
a
21
(t) kf (t)k dt
(t) kg (t)k dt
a
Z b

2
5.3. OTHER REVERSES
Z
219
12
(t) kf (t)k2 dt
(t) kg (t)k2 dt
a
Z b

Z
Z
b
2
(t) kf (t)k dt
12
(t) kg (t)k dt
2
a
b

a
1 (M m)2

4
M +m
The constant
1
4
(t) kg (t)k2 dt.
is best possible.
The case when a function is scalar is incorporated in the following

theorem [3].
Theorem 76 (Dragomir, 2004). Let L2 ([a, b]) , g L2 ([a, b] ; K) ,
and , K with 6= . If e K, kek = 1 and

g (t) + 1

(5.81)
e | |

(t)
2 2
*
+
g (t) g (t)
(5.82)
Re e
,
e 0
(t) (t)
Z
(5.83)
21
(t) | (t)| dt
(t) kg (t)k dt
a
Z b

(t) (t) g (t) dt

Z
0
a
Z
21
(t) | (t)| dt
(t) kg (t)k dt
a
Z b

(t) (t) g (t) dt, e
Z
220
12
(t) | (t)|2 dt
(t) kg (t)k2 dt
a
a

Z b

+

Re
(t) (t) g (t) dt, e
| + |
a
Z b
12
Z b
2
2
(t) | (t)| dt
(t) kg (t)k dt
a
a

Z b

+
Re
(t) (t) g (t) dt, e
| + |
a
Z
1 | |2 b

(t) | (t)|2 dt.
4 | + | a
Z
The constant
1
4
Proof. Follows by Theorem 74 on choosing

v :=
+
e
2
and
r :=
1
| | .
2

Corollary 53. Let L2 ([a, b]) , g L2 ([a, b] ; K) , and M
m > 0. If e K, kek = 1 and

g (t) M + m 1

e (M m)

(t)
2
2
+
*
g (t) g (t)
me 0
,
Re M e
(t) (t)
for a.e. t [a, b], then we have the inequalities:
Z b
12
Z b
2
2
(5.84)
0
(t) | (t)| dt
(t) kg (t)k dt
a
a
Z b

(t) (t) g (t) dt

a
Z
12
(t) | (t)| dt
(t) kg (t)k dt
a
Z b

(t) (t) g (t) dt, e
Z
5.3. OTHER REVERSES
Z
221
21
(t) kg (t)k2 dt
(t) | (t)|2 dt
a
Z b

Re
(t) (t) g (t) dt, e
Z
12
(t) kg (t)k dt
2
(t) | (t)| dt
a
Z b

Re
(t) (t) g (t) dt, e
a
2
The constant
1
4
1 (M m)
4
M +m
(t) | (t)|2 dt.
5.3.3. Applications for the Heisenberg Inequality. The following reverse of the Heisenberg type inequality (5.53) holds [3].
Theorem 77 (Dragomir, 2004). Assume that : [a, b] H is as
in the hypothesis of Theorem 70. In addition, if there exists a r > 0
such that
(5.85)
k0 (t) + t (t)k r

Z b
12
Z b
Z
1 b
2
2
2
0
t k (t)k dt
k (t)k dt
(5.86) 0
k (t)k2 dt
2 a
a
a
1
r2 (b a) .
2
Proof. We observe, by the identity (5.54), that
Z b
Z
1 b
0
(5.87)
Re h (t) , (t) (t)i dt =
k (t)k2 dt.
2
a
a
Now, if we apply Theorem 73 for the choices f (t) = t (t) , g (t) =
1
, t [a, b] , then we deduce the desired inequality
t0 (t) , (t) = ba
(5.86).
Remark 62. It is interesting to remark that, from (5.87), we obviously have
Z b

Z

1 b
2
0

(5.88)
k (t)k dt =
2 a
a
222
Now, if we apply the inequality (see (5.64))

Z b

Z b
Z b

1
2
2

kf (t)k dt
kg (t)k dt
Re hf (t) , g (t)i dt r2 (b a) ,
2
a
a
a
for the choices f (t) = 0 (t) , g (t) = t (t) , t [a, b] , then we get the
same inequality (5.86), but under the condition
k0 (t) t (t)k r
(5.89)
for a.e. t [a, b] .
The following result holds as well [3].

Theorem 78 (Dragomir, 2004). Assume that : [a, b] H is as in
the hypothesis of Theorem 77. In addition, if there exists M m > 0
such that
(5.90)
Re hM t (t) 0 (t) , 0 (t) mt (t)i 0

0
1
M
+
m
(t)
(M m) |t| k (t)k
(5.91)
t
(t)

2
2
Z b
12
Z b
Z
1 b
2
2
2
0
t k (t)k dt
k (t)k dt
(5.92) 0
k (t)k2 dt
2
a
a
a
2 Z b
1 (M m)

t2 k (t)k2 dt.
4
M +m a
Proof. The proof follows by Corollary 52 applied for the function
g (t) = t (t) and f (t) = 0 (t) , and on making use of the identity
(5.88). We omit the details.
Remark 63. If one is interested in reverses for the Heisenberg inequality for real or complex valued functions, then all the other inequalities obtained above for one scalar and one vectorial function may be
applied as well. For the sake of brevity, we do not list them here.
Bibliography
[1] S.S. DRAGOMIR, Reverses of the Cauchy-Bunyakovsky-Schwarz and Heisenberg integral inequalities for vector-valued functions in Hilbert Spaces,
Preprint, RGMIA Res. Rep. Coll., 7(2004), Supplement, Article 20. [Online
http://http://rgmia.vu.edu.au/v7(E).html].
[2] S.S. DRAGOMIR, New reverses of Schwarz, triangle and Bessel inequalities in
inner product spaces, Australian J. Math. Anal. & Appl., 1(2004), No.1, Article
1 [Online http://ajmaa.org].
[3] S.S. DRAGOMIR, New reverses of the Cauchy-Bunyakovsky-Schwarz integral inequality for vector-valued functions in Hilbert spaces and applications,
Preprint, RGMIA Res. Rep. Coll., 7(2004), Supplement, Article 21. [Online
http://http://rgmia.vu.edu.au/v7(E).html].
[4] G.H. HARDY, J.E. LITTLEWOOD and G. POLYA, Inequalities, Cambridge
University Press, Cambridge, United Kingdom, 1952.
223
CHAPTER 6
Other Inequalities in Inner Product Spaces

6.1. Bounds for the Distance to Finite-Dimensional
Subspaces
6.1.1. Introduction. Let (H; h, i) be an inner product space
over the real or complex number field K, {y1 , . . . , yn } a subset of H
and G (y1 , . . . , yn ) the Gram matrix of {y1 , . . . , yn } where (i, j) entry
is hyi , yj i . The determinant of G (y1 , . . . , yn ) is called the Gram determinant of {y1 , . . . , yn } and is denoted by (y1 , . . . , yn ) . Thus,

hy1 , y1 i hy1 , y2 i hy1 , yn i

hy2 , y1 i hy2 , y2 i hy2 , yn i
.
(y1 , . . . , yn ) =

hyn , y1 i hyn , y2 i hyn , yn i
Following [4, p. 129 133], we state here some general results for
the Gram determinant that will be used in the sequel.
(1) Let {x1 , . . . , xn } H. Then (x1 , . . . , xn ) 6= 0 if and only if
{x1 , . . . , xn } is linearly independent;
(2) Let M = span {x1 , . . . , xn } be ndimensional in H, i.e., {x1 , . . . ,
xn } is linearly independent. Then for each x H, the distance
d (x, M ) from x to the linear subspace H has the representations
d2 (x, M ) =
(6.1)
(x1 , . . . , xn , x)
(x1 , . . . , xn )
and
(6.2)
d2 (x, M ) = kxk2 T G1 ,
where G = G (x1 , . . . , xn ) , G1 is the inverse matrix of G and
T = (hx, x1 i , hx, x2 i , . . . , hx, xn i) ,
denotes the transpose of the column vector .
225
226
6. OTHER INEQUALITIES
(6.3)
Moreover, one has the simpler representation
P
2
( ni=1 |hx,xi i|2 )
if x
/ M ,
kxk Pn
2
hx,x
ix
k
k
i
i
i=1
d2 (x, M ) =
if x M ,
kxk2
where M denotes the orthogonal complement of M.

(3) Let {x1 , . . . , xn } be a set of nonzero vectors in H. Then
0 (x1 , . . . , xn ) kx1 k2 kx2 k2 kxn k2 .
(6.4)
The equality holds on the left (respectively right) side of (6.4)

if and only if {x1 , . . . , xn } is linearly dependent (respectively
orthogonal). The first inequality in (6.4) is known in the literature as Grams inequality while the second one is known as
Hadamards inequality.
(4) If {x1 , . . . , xn } is an orthonormal set in H, i.e., hxi , xj i = ij ,
i, j {1, . . . , n} , where ij is Kroneckers delta, then
2
d (x, M ) = kxk
(6.5)
n
X
|hx, xi i|2 .
i=1
The following inequalities which involve Gram determinants may

be stated as well [17, p. 597]:
(6.6)
(x2 , . . . , xn )
(x1 , . . . , xn )
(xk+1 , . . . , xn ) ,
(x1 , . . . , xk )
(x1 , . . . , xk )
(x1 , . . . , xn ) (x1 , . . . , xk ) (xk+1 , . . . , xn )
(6.7)
and
(6.8)
2 (x1 + y1 , x2 , . . . , xn )
1
2 (x1 , x2 , . . . , xn ) + 2 (y1 , x2 , . . . , xn ) .
The main aim of this section is to point out some upper bounds
for the distance d (x, M ) in terms of the linearly independent vectors
{x1 , . . . , xn } that span M and x
/ M , where M is the orthogonal
complement of M in the inner product space (H; h, i).
As a by-product of this endeavour, some refinements of the generalisations for Bessels inequality due to several authors including: Boas,
Bellman and Bombieri are obtained. Refinements for the well known
Hadamards inequality for Gram determinants are also derived.
6.1. BOUNDS FOR THE DISTANCE TO FINITE-DIMENSIONAL SUBSPACES

227
6.1.2. Upper Bounds for d (x, M ). The following result may be

stated [16].
Theorem 79 (Dragomir, 2005). Let {x1 , . . . , xn } be a linearly independent system of vectors in H and M := span {x1 , . . . , xn } . If
x
/ M , then
P
P
kxk2 ni=1 kxi k2 ni=1 |hx, xi i|2
2
(6.9)
d (x, M ) <
Pn
2
i=1 kxi k
or, equivalently,
(6.10) (x1 , . . . , xn , x)
kxk2
<
Pn
i=1
P
kxi k2 ni=1 |hx, xi i|2
(x1 , . . . , xn ) .
Pn
2
kx
k
i
i=1
Proof. If we use the Cauchy-Bunyakovsky-Schwarz type inequality

n
2
n
n
X

X
X

2
(6.11)
i yi
|i |
kyi k2 ,

i=1
i=1
i=1
that can be easily deduced from the obvious identity

n
2
n
n
n
X

X
X
1X

2
2
(6.12)
ki xj j xi k2 ,
|i |
kyi k
i yi =

2
i=1
i=1
i=1
i,j=1
we can state that

2
n
n
n
X

X
X

2
(6.13)
hx, xi i xi
|hx, xi i|
kxi k2 .

i=1
i=1
i=1
Note that the equality case holds in (6.13) if and only if, by (6.12),
(6.14)
hx, xi ixj = hx, xi ixi
for each i, j {1, . . . , n} .

Utilising the expression (6.3) of the distance d (x, M ), we have
Pn
2 Pn
2 Pn
2
2
2
i=1 |hx, xi i|
i=1 |hx, xi i|
i=1 kxi k
(6.15) d (x, M ) = kxk
P
Pn
2
n
2 .
kx
k
k i=1 hx, xi i xi k
i
i=1
Since {x1 , . . . , xn } are linearly independent, hence (6.14) cannot be
achieved and then we have strict inequality in (6.13).
Finally, on using (6.13) and (6.15) we get the desired result (6.9).
228
Remark 64. It is known that (see (6.4)) if not all {x1 , . . . , xn } are
orthogonal on each other, then the following result, which is well known
in the literature as Hadamards inequality holds:
(x1 , . . . , xn ) < kx1 k2 kx2 k2 kxn k2 .
(6.16)
Utilising the inequality (6.10), we may write successively:

kx1 k2 kx2 k2 |hx2 , x1 i|2
(x1 , x2 )
kx1 k2 kx1 k2 kx2 k2 ,
2
kx1 k
P2
2 P2
2
2
kx3 k
i=1 kxi k
i=1 |hx3 , xi i|
(x1 , x2 , x3 ) <
(x1 , x2 )
P2
2
i=1 kxi k
kx3 k2 (x1 , x2 )
P
Pn1
2
2
kxn k2 n1
i=1 kxi k
i=1 |hxn , xi i|
(x1 , . . . , xn1 , xn ) <
Pn1
2
i=1 kxi k
(x1 , . . . , xn1 )
kxn k2 (x1 , . . . , xn1 ) .
Multiplying the above inequalities, we deduce
(6.17)
(x1 , . . . , xn1 , xn )
< kx1 k
n
Y
kxk k Pk1
i=1
k=2
n
Y
kxi k2
k1
X
!
2
|hxk , xi i|
i=1
kxj k2 ,
j=1
valid for a system of n 2 linearly independent vectors which are not

orthogonal on each other.
In [15], the author has obtained the following inequality.
Lemma 8 (Dragomir, 2004). Let z1 , . . . , zn H and 1 , . . . , n
K. Then one has the inequalities:
(6.18)

2
n
X

i zi

i=1

229
n
P
max |i |2
kzi k2 ;
1in
i=1
n
1 n
p1
P
P
2
2
kzi k
|i |
i=1
i=1
where > 1, 1 + 1 = 1;
|i |2 max kzi k2 ;
1in
i=1
P
max {|i j |}
|hzi , zj i| ;
1i6=jn
1i6=jn
"
# 1
! 1
2
n
n
P
P
P
|i |
|i |2
|hzi , zj i|
i=1
i=1
1i6=jn
+
where
> 1, 1 + 1 = 1;
"
#
2
n
n
P
P
|i |
|i |2 max |hzi , zj i| ;
1i6=jn
i=1
i=1
where any term in the first branch can be combined with each term from
the second branch giving 9 possible combinations.
Out of these, we select the following ones that are of relevance for
further consideration:
(6.19)

2
n
X

i zi

i=1
max kzi k
1in
n
X
|i |2
i=1
+ max |hzi , zj i|
1i<jn
n
X
i=1
|i |
n
X
!2
|i |
i=1
n
X
|i |2
i=1

max kzi k + (n 1) max |hzi , zj i|
2
1in
1i<jn
230
and

2
n
X

i zi

(6.20)
i=1
max kzi k
n
X
2
1in
|i |2 +
n
X
i=1
!2
|i |2
i=1
n
X
1/2
|i |4
i=1
! 21
X
|hzi , zj i|
1i6=jn
n
X
! 12
|i |2 max kzi k2 +
1in
i=1
|hzi , zj i|2
1i6=jn
Note that the last inequality in (6.19) follows by the fact that
n
X
!2
|i |
i=1
n
X
|i |2 ,
i=1
while the last inequality in (6.20) is obvious.

Utilising the above inequalities (6.19) and (6.20) which provide alternatives to the Cauchy-Bunyakovsky-Schwarz inequality (6.11), we
can state the following results [16].
Theorem 80 (Dragomir, 2005). Let {x1 , . . . , xn } , M and x be as
in Theorem 79. Then
(6.21) d2 (x, M )
! 12
kxk2 max kxi k2 +

1in
|hxi , xj i|2
i=1
1i6=jn
! 21
max kxi k2 +
1in
P
1i6=jn
n
P
|hxi , xj i|2
|hx, xi i|2

231
or, equivalently,
(6.22) (x1 , . . . , xn , x)
! 12
kxk2 max kxi k2 +
|hxi , xj i|2
1in
!
P
max kxi k +
1in
|hx, xi i|2
i=1
1i6=jn
n
P
1
2
|hxi , xj i|
1i6=jn
(x1 , . . . , xn ) .
Proof. Utilising the inequality (6.20) for i = hx, xi i and zi = xi ,
i {1, . . . , n} , we can write:

2
n
X

(6.23)
hx, xi i xi

i=1
n
X
i=1
! 12
|hx, xi i|2 max kxi k2 +

1in
|hxi , xj i|2
1i6=jn
for any x H.
Now, since, by the representation formula (6.3)
Pn
n
2
X
2
2
i=1 |hx, xi i|
(6.24)
d (x, M ) = kxk Pn
|hx, xi i|2 ,
2
k i=1 hx, xi i xi k i=1
for x
/ M , hence, by (6.23) and (6.24) we deduce the desired result
(6.21).
Remark 65. In 1941, R.P. Boas [2] and in 1944, R. Bellman
[1], independent of each other, proved the following generalisation of
Bessels inequality:
! 21
n
X
X
(6.25)
|hy, yi i|2 kyk2 max kyi k2 +
|hyi , yj i|2 ,
i=1
1in
1i6=jn
provided y and yi (i {1, . . . , n}) are arbitrary vectors in the inner

product space (H; h, i) . If {yi }i{1,...,n} are orthonormal, then (6.25)
reduces to Bessels inequality.
In this respect, one may see (6.21) as a refinement of the BoasBellman result (6.25).
232
Remark 66. On making use of a similar argument to that utilised

in Remark 64, one can obtain the following refinement of the Hadamard
inequality:
(6.26)
(x1 , . . . , xn )
kx1 k2
k1
P
kxk k
k=2
n
Y
i=1
max kxi k2 +
1ik1
n
Y
|hxk , xi i|2
P
! 12
2
|hxi , xj i|
1i6=jk1
kxj k2 .
j=1
Further on, if we choose i = hx, xi i , zi = xi , i {1, . . . , n} in

(6.19), then we may state the inequality
n
2
X

(6.27)
hx, xi i xi

i=1
n
X
|hx, xi i|
i=1

max kxi k + (n 1) max |hxi , xj i| .
2
1in
1i6=jn
Utilising (6.27) and (6.24) we may state the following result as well
[16]:
in Theorem 79. Then
(6.28) d2 (x, M )

P
2
2
kxk max kxi k + (n 1) max |hxi , xj i| ni=1 |hx, xi i|2
1in
1i6=jn
2
max kxi k + (n 1) max |hxi , xj i|
1in
1i6=jn

233
or, equivalently,
(6.29) (x1 , . . . , xn , x)

P
2
2
kxk max kxi k + (n 1) max |hxi , xj i| ni=1 |hx, xi i|2
1in
1i6=jn
max kxi k2 + (n 1) max |hxi , xj i|

1in
1i6=jn
(x1 , . . . , xn ) .
Remark 67. The above result (6.28) provides a refinement for the
following generalisation of Bessels inequality:
(6.30)
n
X
|hx, xi i| kxk
i=1

max kxi k + (n 1) max |hxi , xj i| ,
2
1in
1i6=jn
obtained by the author in [15].

One can also provide the corresponding refinement of Hadamards
inequality (6.4) on using (6.29), i.e.,
(6.31) (x1 , . . . , xn )
kx1 k2
k1
P
|hxk , xi i|
n
Y
i=1
2
kxk k
max kxi k2 + (k 2) max

k=2
1ik1
n
Y
1i6=jk1
|hxi , xj i|
kxj k2 .
j=1
6.1.3. Other Upper Bounds for d (x, M ). In [7, p. 140] the

author obtained the following inequality that is similar to the CauchyBunyakovsky-Schwarz result.
234

K. Then one has the inequalities:

2
n
n
n
X

X
X

2
(6.32)
i zi
|i |
|hzi , zj i|

i=1
i=1
j=1
"
#
n
n
P
P
|i |2 max
|hzi , zj i| ;
1in j=1
i=1
!q ! 1q
p1 n
P
n
n
P
P
|hzi , zj i|
|i |2p
i=1 j=1
i=1
where p > 1, p1 + 1q = 1;
max |i |2
|hzi , zj i| .
1in
i,j=1
We can state and prove now another upper bound for the distance
d (x, M ) as follows [16].
in Theorem 79. Then
"
#
n
n
P
P
kxk2 max
|hxi , xj i|
|hx, xi i|2
1in
d2 (x, M )
(6.33)
j=1
i=1
"
max
1in
n
P
#
|hxi , xj i|
j=1
or, equivalently,
(6.34) (x1 , . . . , xn , x)
"
2
kxk max
1in
n
P
#
|hxi , xj i|
j=1
i=1
"
max
1in
n
P
n
P
|hx, xi i|2
(x1 , . . . , xn ) .
|hxi , xj i|
j=1
Proof. Utilising the first branch in (6.32) we may state that

2
" n
#
n
n
X

X
X

(6.35)
hx, xi i xi
|hx, xi i|2 max
|hxi , xj i|

1in

i=1
for any x H.
i=1
j=1

235
Now, since, by the representation formula (6.3) we have

Pn
n
2
X
2
2
i=1 |hx, xi i|
(6.36)
d (x, M ) = kxk Pn
|hx, xi i|2 ,
2
for x
/ M , hence, by (6.35) and (6.36) we deduce the desired result
(6.33).
Remark 68. In 1971, E. Bombieri [3] proved the following generalisation of Bessels inequality, however not stated in the general form
for inner products. The general version can be found for instance in
[17, p. 394]. It reads as follows: if y, y1 , . . . , yn are vectors in the inner
product space (H; h, i) , then
( n
)
n
X
X
2
2
(6.37)
|hy, yi i| kyk max
|hyi , yj i| .
1in
i=1
j=1
Obviously, when {y1 , . . . , yn } are orthonormal, the inequality (6.37)

produces Bessels inequality.
In this respect, we may regard our result (6.33) as a refinement of
the Bombieri inequality (6.37).
Remark 69. On making use of a similar argument to that in Remark 64, we obtain the following refinement for the Hadamard inequality:
k1
P
|hxk , xi i|2
n
Y
i=1
2
2
#
"
(6.38) (x1 , . . . , xn ) kx1 k
kxk k
k1
P
k=2
max
|hxi , xj i|
1ik1
n
Y
j=1
kxj k2 .
j=1
Another different Cauchy-Bunyakovsky-Schwarz type inequality is

incorporated in the following lemma [13].
K. Then

2
! p2
! 1q
n
n
n
X

X
X

(6.39)
i zi
|i |p
|hzi , zj i|q

i=1
for p > 1,
1
p
1
q
i=1
= 1.
i,j=1
236
If in (6.39) we choose p = q = 2, then we get

2
! 12
n
n
n
X

X
X

.
(6.40)
i zi
|i |2
|hzi , zj i|2

i=1
i=1
i,j=1
Based on (6.40), we can state the following result that provides yet
another upper bound for the distance d (x, M ) [16].
in Theorem 79. Then
! 21
n
n
P
P
kxk2
|hxi , xj i|2
|hx, xi i|2
(6.41)
i=1
i,j=1
d2 (x, M )
! 12
n
P
|hxi , xj i|2
i,j=1
or, equivalently,
(6.42) (x1 , . . . , xn , x)
kxk2
n
P
! 12
|hxi , xj i|2
i,j=1
n
P
i=1
n
P
! 12
|hx, xi i|2
(x1 , . . . , xn ) .
|hxi , xj i|2
i,j=1
Similar comments apply related to Hadamards inequality. We omit

the details.
6.1.4. Some Conditional Bounds. In the recent paper [6], the
author has established the following reverse of the Bessel inequality.
Let (H; h, i) be an inner product space over the real or complex
number field K, {ei }iI a finite family of orthonormal vectors in H,
i , i K, i I and x H. If
*
+
X
X
(6.43)
Re
i ei x, x
i ei 0
iI
or, equivalently,

X +
1

i
i
ei
(6.44)
x

2
2
iI
iI
! 12
X
iI
|i i |2

237
then
(6.45)
(0 ) kxk2
X
iI
|hx, ei i|2
1X
| i |2 .
4 iI i
1
4
The constant is best possible in the sense that it cannot be replaced

by a smaller constant [16].
Theorem 84 (Dragomir, 2005). Let {x1 , . . . xn } be a linearly independent system of vectors in H and M := span {x1 , . . . xn } . If i ,
i K, i {1, . . . , n} and x H\M is such that
* n
+
n
X
X
(6.46)
Re
i xi x, x
i xi 0,
i=1
i=1
then we have the bound

2
n

X
1

2
(i i ) xi
d (x, M )

4 i=1
(6.47)
or, equivalently,
(6.48)
n
2

1
X

(x1 , . . . , xn , x)
(i i ) xi (x1 , . . . , xn ) .

4 i=1
Proof. It is easy to see that in an inner product space for any

x, z, Z H one has
2

x z + Z 1 kZ zk2 = Re hZ x, x zi ,

2
4
therefore, the condition (6.46) is actually equivalent to

2
n
2
n

X
i + i
1

X

(6.49)
xi
(i i ) xi .
x

2
4 i=1
i=1
Now, obviously,
(6.50)

2
n

X

i
i
d2 (x, M ) = inf kx yk2 x
xi
yM

2
i=1
and thus, by (6.49) and (6.50) we deduce (6.47).

The last inequality is obvious by the representation (6.2).
Remark 70. Utilising various Cauchy-Bunyakovsky-Schwarz type
inequalities we may obtain more convenient (although coarser) bounds
238
for d2 (x, M ) . For instance, if we use the inequality (6.19) we can state
the inequality:
n
2
X

(i i ) xi

i=1

n
X
2
2
|i i | max kxi k + (n 1) max |hxi , xj i| ,

1in
i=1
1i<jn
giving the bound:

n
1X
(6.51) d (x, M )
|i i |2
4 i=1

2
max kxi k + (n 1) max |hxi , xj i| ,
2
1in
1i<jn

Obviously, if {x1 , . . . , xn } is an orthonormal family in H, then from
(6.51) we deduce the reverse of Bessels inequality incorporated in (6.45).
If we use the inequality (6.20), then we can state the inequality

2
n
X

(
)
x

i
i
i

i=1
! 12
n
X
X
|i i |2 max kxi k2 +
|hxi , xj i|2 ,
1in
i=1
1i6=jn
giving the bound

n
1X
(6.52) d (x, M )
|i i |2
4 i=1
max kxi k2 +
1in
! 12
X
|hxi , xj i|2
1i6=jn

In this case, when one assumes that {x1 , . . . , xn } is an orthonormal
family of vectors, then (6.52) reduces to (6.45) as well.
Finally, on utilising the first branch of the inequality (6.32), we can
state that
" n
#
n
X
X
1
2
(6.53)
d2 (x, M )
|i i | max
|hxi , xj i| ,
1in
4 i=1
j=1
6.2. REVERSING THE CBS INEQUALITY FOR SEQUENCES
239

This inequality is also a generalisation of (6.45).
6.2. Reversing the CBS Inequality for Sequences
over the real or complex number field K. One of the most important inequalities in inner product spaces with numerous applications,
is the Schwarz inequality
|hx, yi|2 kxk2 kyk2 ,
(6.54)
x, y H
or, equivalently,
|hx, yi| kxk kyk ,
(6.55)
x, y H.
The case of equality holds iff there exists a scalar K such that
x = y.
By a multiplicative reverse of the Schwarz inequality we understand
an inequality of the form
(6.56)
(1 )
kxk kyk
kxk2 kyk2
k1 or (1 )
k2
|hx, yi|
|hx, yi|2
with appropriate k1 and k2 and under various assumptions for the vectors x and y, while by an additive reverse we understand an inequality
of the form
(6.57)
(0 ) kxk kyk |hx, yi| h1 or

(0 ) kxk2 kyk2 |hx, yi|2 h2 .
Similar definition apply when |hx, yi| is replaced by Re hx, yi or

|Re hx, yi| .
The following recent reverses for the Schwarz inequality hold (see
for instance the monograph on line [7, p. 20]).
space over the real or complex number field K. If x, y H and r > 0
are such that
kx yk r < kyk ,
(6.58)
then we have the following multiplicative reverse of the Schwarz inequality

(6.59)
(1 )
kxk kyk
kxk kyk
kyk
q
|hx, yi|
Re hx, yi
kyk2 r2
240
and the subsequent additive reverses

(6.60)
r2
kyk

r2
kyk +
kyk
Re hx, yi
r2
and
(6.61)
kxk2 kyk2 [Re hx, yi]2

r2 kxk2 .
All the above inequalities are sharp.
Other additive reverses of the quadratic Schwarzs inequality are
incorporated in the following result [7, p. 18-19].
Theorem 86 (Dragomir, 2004). Let x, y H and a, A K. If
Re hAy x, x ayi 0
(6.62)
or, equivalently,

1

a
+
A
|A a| kyk ,
x
y
2

2
(6.63)
then
(6.64)

A+a kyk2 hx, yi2
2
1
|A a|2 kyk4
4
kyk2 Re hAy x, x ayi
1
|A a|2 kyk4 .
4
The constant
1
4
is best possible in all inequalities.
If one were to assume more about the complex numbers A and a,

then one may state the following result as well [7, p. 21-23].
Theorem 87 (Dragomir, 2004). With the assumptions of Theorem
86 and, if in addition, Re (A
a) > 0, then

hx, yi
1
1 Re A + a
|A + a|
p
p
|hx, yi| ,
(6.65) kxk kyk
2
2
Re (A
a)
Re (A
a)
(6.66)
241

h

i
p
2 Re (A
a) hx, yi
1 Re A + a
p

2
Re (A
a)
and
(6.67)
1 |A a|2
|hx, yi|2 .
4 Re (A
a)
2
The constants
1
2
and
1
4
are best possible.
Remark 71. If A = M, a = m and M m > 0, then (6.65) and

(6.66) may be written in a more convenient form as
M +m
kxk kyk
Re hx, yi
2 mM
(6.68)
and

(6.69)
Here the constant
1
2
2
M m
Re hx, yi .
2 mM
is sharp in both inequalities.
In this section several reverses for the Cauchy-Bunyakovsky-Schwarz

(CBS) inequality for sequences of vectors in Hilbert spaces are obtained.
Applications for bounding the distance to a finite-dimensional subspace
and in reversing the generalised triangle inequality are also given.
6.2.2. Reverses of the (CBS) Inequality for Two Sequences
in `2 (K). Let (K, h, i) be a Hilbert space over K, pi 0, i N with
Pp
2
i=1 pi = 1. Consider `p (K) as the space

(
)

X

2
2
`p (K) := x = (xi )iN xi K, i N and
pi kxi k < .

i=1
It is well known that `2p (K) endowed with the inner product
hx, yip :=
pi hxi , yi i
i=1
is a Hilbert space over K. The norm kkp of `2p (K) is given by

! 12
X
kxkp :=
pi kxi k2
.
i=1
242
If x, y `2p (K) , then the following Cauchy-Bunyakovsky-Schwarz (CBS)

inequality holds true
(6.70)
pi kxi k
X
2
i=1
i=1
X

pi kyi k2
pi hxi , yi i

i=1
with equality iff there exists a K such that xi = yi for each i N.

This is an obvious consequence of the Schwarz inequality (6.54)
written for the inner product h, ip defined on `2p (K) .
The following proposition may be stated [11].
Proposition 57. Let x, y `2p (K) and r > 0. Assume that
kxi yi k r < kyi k for each i N.
(6.71)

1
P
2 2
pi kxi k2
i=1 pi kyi k
P
(1 )
|
i=1 pi hxi , yi i|
1
P
2 P
2 2
i=1 pi kxi k
i=1 pi kyi k
P
i=1 pi Re hxi , yi i
1
P
2 2
i=1 pi kyi k
qP
,
2
2
p
ky
k
r
i
i=1 i
i=1
(6.72)
(6.73) (0 )
pi kxi k
X
2
i=1
X
i=1
! 12
pi kyi k2
i=1
i=1
pi kxi k2

X

pi hxi , yi i

! 12
pi kyi k2
pi Re hxi , yi i
i=1
i=1
r2
pi Re hxi , yi i
i=1
"
#
r
12 r
P
P
P
pi kyi k2 r2
+
pi kyi k2 r2
pi kyi k2
i=1
i=1
i=1
243
and

2
X

2
(0 )
pi kxi k
pi kyi k
pi hxi , yi i

i=1
i=1
i=1
"
#2
X
X
X
pi kxi k2
pi kyi k2
pi Re hxi , yi i
(6.74)
i=1
r2
X
2
i=1
i=1
pi kxi k2 .
i=1

kx
yk2p
X
i=1
pi kxi yi k r
pi
i=1
pi kyi k2 = kyk2p ,
i=1
giving kx ykp r kykp . Applying Theorem 85 for `2p (K) and

h, ip , we deduce the desired inequality.
The following proposition holds [11].
Proposition 58. Let x, y `2p (K) and a, A K. If
(6.75)
Re hAyi xi , xi ayi i 0 for each i N
or, equivalently,

1
a
+
A
xi
(6.76)
yi
2 |A a| kyi k

2
for each i N
then
(6.77)
X

X

2
2
(0 )
pi kxi k
pi kyi k
pi hxi , yi i

i=1
i=1
i=1
!
2
X
1
2
2
pi kyi k
|A a|
4
i=1
A+a P
2
P
2

i=1 pi kyi k
i=1 pi hxi , yi i
2
P
2 P
i=1 pi kyi k
i=1 pi Re hAyi xi , xi ayi i
!2
X
1
|A a|2
pi kyi k2 .
4
i=1
The proof follows by Theorem 86, we omit the details.

Finally, on using Theorem 87, we may state [11]:
244
Proposition 59. Assume that x, y, a and A are as in Proposition

58. Moreover, if Re (A
a) > 0, then we have the inequality:
! 12
X
X
(6.78)
pi kxi k2
pi kyi k2
i=1
(6.79)
i=1
(0 )
P

Re A + a
p
hx
,
y
i
i
i
i
i=1
p
Re (A
a)

|A a| X

p
pi hxi , yi i ,

Re (A
a)
i=1
pi kxi k
i=1
1 Re
X
2
! 21
pi kyi k2
pi Re hxi , yi i
i=1
i=1
h
P
i
p
A+a
2 Re (A
a)
i=1 pi hxi , yi i
p
Re (A
a)
and
(6.80)
X

(0 )
pi kxi k
pi kyi k
pi hxi , yi i

i=1
i=1
i=1

2

1 |A a|2 X

p
hx
,
y
i

i
i i .

4 Re (A
a) i=1
6.2.3. Reverses of the (CBS) Inequality for Mixed Sequences. Let

space over K and for pi 0,
P(K, h, i) be a Hilbert
2
i N with i=1 pi = 1, and `p (K) the Hilbert space defined in the
previous section.
If

)
(

X

2
`2p (K) := = (i )iN i K, i N and
pi |i | <

i=1
and x `2p (K) , then the following Cauchy-Bunyakovsky-Schwarz (CBS)

inequality holds true:

2
X

X
X

2
2
(6.81)
pi |i |
pi kxi k
pi i xi ,

i=1
i=1
i=1
with equality if and only if there exists a vector v K such that

xi = i v for any i N.
245
The inequality (6.81) follows by the obvious identity

n
X
i=1

2
n
n
X

X

pi |i |2
pi kxi k2
pi i xi

i=1
i=1
1 XX
=
pi pj ki xj j xi k2 ,
2 i=1 j=1
for any n N, n 1.
In the following we establish some reverses of the (CBS) inequality
in some of its various equivalent forms that will be specified where they
occur [11].
Theorem 88 (Dragomir, 2005). Let `2p (K) , x `2p (K) and
a K, r > 0 such that kak > r. If the following condition holds
kxi i ak r |i |
(6.82)
for each i N,
(note that if i 6= 0 for any i N, then the condition (6.82) is equivalent to

xi

r for each i N),
(6.83)
a
i

then we have the following inequalities
(6.84)
X
2
pi |i |
pi kxi k2
i=1
i=1
! 12
kak2 r2
kak
kak2 r2
(6.85)
Re
*
X
! 21
+
pi i xi , a
i=1

X

pi i xi ;

i=1

X

pi i xi

i=1
i=1
i=1
*
! 12
+
X
X
X
a
pi |i |2
pi kxi k2
Re
pi i xi ,
kak
i=1
i=1
i=1
X
2
pi |i |
pi kxi k2
246
r2
kak
kak +
Re
kak
r2
r2
kak
(6.86)

r2

r2
kak +
*
X
kak2 r
i=1
a
pi i xi ,
kak
X

pi i xi ;

2
i=1
" *
+#2
X
1
Re
pi i xi , a
kak2 r2
i=1

2
2

X
kak

pi i xi

2
2

kak r
X
2
pi |i |
pi kxi k2
i=1
i=1
i=1
and

2
X

0
pi |i |
pi kxi k
pi i xi

i=1
i=1
i=1
"
*
+#2
X
X
X
a
pi |i |2
pi kxi k2 Re
pi i xi ,
kak
i=1
i=1
i=1
" *
+#2
X
r2
Re
pi i xi , a
kak2 kak2 r2
i=1

2

2
X
r

pi i xi .

2
2

kak r
(6.87)
i=1
All the inequalities in (6.84) (6.87) are sharp.

kxi k2 2 Re hxi , i ai + |i |2 kak2 |i |2 r2
for any i N, which is clearly equivalent to

(6.88)
kxi k2 + kak2 r2 |i |2 2 Re hi xi , ai
for each i N.
If we multiply (6.88) by pi 0 and sum over i N, then we deduce
*
+
X
X
X

(6.89)
pi kxi k2 + kak2 r2
pi |i |2 2 Re
pi i xi , a .
i=1
i=1
i=1
Now, dividing (6.89) by
(6.90) q
kak2 r2
247
kak2 r2 > 0 we get

2
pi kxi k +
kak
r2
i=1
pi |i |2
i=1
2
2
Re
kak
r2
*
X
+
pi i xi , a .
i=1

1
p + q 2 pq,
> 0, p, q 0,
we can state that:

"
# 21
X
X
(6.91) 2
pi |i |2
pi kxi k2
i=1
i=1
kak r2
pi kxi k +
kak
r2
i=1
pi |i |2 .
i=1
Making use of (6.90) and (6.91), we deduce the first part of (6.84).
The second part is obvious by Schwarzs inequality

*
+
X

X

Re
pi i xi , a
pi i xi kak .

i=1
i=1
If p1 = 1, x1 = x, 1 = 1 and pi = 0, i = 0, xi = 0 for i 2, then

from (6.84) we deduce the inequality
kxk kak
Re hx, ai q
kak2 r2
kak2 r2
kxk q
provided kx ak r < kak , x, a K. The sharpness of this inequality

has been shown in [7, p. 20], and we omit the details.
The other inequalities are obvious consequences of (6.84) and we
omit the details.
The following corollary may be stated [11].
Corollary 54. Let `2p (K) , x `2p (K) , e H, kek = 1 and
, K with Re (
) > 0. If

+

1
(6.92)
xi i 2 e 2 | | |i |
248
for each i N, or, equivalently

Re hi e xi , xi i ei 0
(6.93)
for each i N, (note that, if i 6= 0 for any i N, then (6.92) is

equivalent to

1
xi
| |

e
(6.94)
2
i
2
for each i N and (6.93) is equivalent to

xi xi
Re e ,
e 0
i i
for each i N), then the following reverses of the (CBS) inequality
are valid:
! 21

P
+
X
X
Re
h
p
x
,
ei
i
i
i
i=1
pi |i |2
pi kxi k2
(6.95)
1
2 [Re ()] 2
i=1
i=1

1
| + |
X

p
x

i i i ;
1

2 [Re ()] 2 i=1
(6.96)
X
2
pi |i |
pi kxi k2
i=1
i=1
pi |i |
i=1
! 12

X

pi i xi

i=1
! 12
pi kxi k2
i=1
"
Re
| + |
*
X
+#
pi i xi , e
i=1
2
| |

p
p
2 Re () | + | + 2 Re ()
"
*
+#
+
X
Re
pi i xi , e
| + | i=1

X

p
pi i xi ;
p

2 Re () | + | + 2 Re ()
| |2
i=1
(6.97)
pi |i |2
i=1
249
pi kxi k2
i=1
" (
*
+)#2
X
1
+
Re
pi i xi , e
4 Re (
)
i=1

2
2 X

1 | + |

pi i xi

4 Re (
)
i=1
and
(6.98)
X
i=1
pi |i |
pi |i |2
X
i=1
i=1

2
X

pi kxi k
pi i xi

2
i=1
pi kxi k2
i=1
"
Re
| + |
*
X
+)#2
pi i xi , e
i=1
"
*
+#)2
X
| |
+
Re
pi i xi , e
4 | + |2 Re (
)
i=1

2

X
| |2

p

i i i .

4 Re (
)
2
i=1
All the inequalities in (6.95) (6.98) are sharp.

Remark 72. We remark that if M m > 0 and for `2p (K) ,
x `2p (K) , e H with kek = 1, one would assume that either
(6.99)

xi
1
M
+
m

(M m)
e
i
2
2
for each i N, or, equivalently

(6.100)

xi xi
Re M e ,
me 0
i i
250
for each i N, then the following, much simpler reverses of the (CBS)
inequality may be stated:
(6.101)
(6.102)
X
2
pi |i |
pi kxi k2
i=1
i=1
! 12
*
+
X
M +m

Re
pi i xi , e
2 mM
i=1

M +m
X

pi i xi ;

2 mM i=1
! 12

X

pi i xi

i=1
i=1
i=1
! 12
*
+
X
X
X
pi |i |2
pi kxi k2
Re
pi i xi , e
X
pi |i |2
pi kxi k2
i=1
i=1
i=1
2
*
X
(M m)
Re
pi i xi , e
2
i=1
2
M+ m
mM

X

(M m)2

p
x

i i i ;
2

2
M+ m
mM i=1
(6.103)
X

X

pi |i |2
pi kxi k2
pi i xi

i=1
i=1
i=1
" *
+#2
X
(M + m)2
Re
pi i xi , e
4mM
i=1

2

X
(M + m)2

pi i xi

4mM
i=1
and
(6.104)
X
i=1
X

X

pi |i |2
pi kxi k2
pi i xi

i=1
i=1
" *
+#2
X
X
pi |i |2
pi kxi k2 Re
pi i xi , e
i=1
i=1
2
251
(M m)
4mM
(M m)2
4mM
"
Re
i=1
*
X
+#2
pi i xi , e
i=1

2
X

p
x

i i i .

i=1
6.2.4. Reverses for the Generalised Triangle Inequality. In

1966, J.B. Diaz and F.T. Metcalf [5] proved the following reverse of
the generalised triangle inequality holding in an inner product space
(H; h, i) over the real or complex number field K:

n
n
X

X

(6.105)
r
kxi k
xi

i=1
i=1
provided the vectors x1 , . . . , xn H\ {0} satisfy the assumption

(6.106)
0r
Re hxi , ai
,
kxi k
where a H and kak = 1.

In an attempt to diversify the assumptions for which such reverse
results hold, the author pointed out in [10] that

n
n
X

X
p

2
(6.107)
1
kxi k
xi ,

i=1
i=1
where the vectors xi, i {1, . . . , n} satisfy the condition

(6.108)
kxi ak ,
i {1, . . . , n}
where a H, kak = 1 and (0, 1) .

If, for M m > 0, the vectors xi H, i {1, . . . , n} verify either
(6.109)
Re hM a xi , xi mai 0,
or, equivalently,

1
M
+
m
xi
(M m) ,
(6.110)
a
2

2
i {1, . . . , n} ,
i {1, . . . , n} ,
where a H, kak = 1, then the following reverse of the generalised

triangle inequality may be stated as well [10]

n
n
X

2 mM X

kxi k
xi .
(6.111)

M + m i=1
i=1
252
Note that the inequalities (6.105), (6.107), and (6.111) are sharp;
necessary and sufficient equality conditions were provided (see [5] and
[10]).
It is obvious, from Theorem 88, that, if
(6.112)
kxi ak r,
for
i {1, . . . , n} ,
where kak > r, a H and xi H, i {1, . . . , n} , then one can state

the inequalities
! 12
n
n
X
X
(6.113)
kxi k n
kxi k2
i=1
i=1
Re
* n
X
kak2 r2
kak
kak2 r2
+
xi , a
i=1

n
X

xi

i=1
and
(6.114)
n

X

0
kxi k
xi

i=1
i=1

! 12 n
n
X
X

2
kxi k

xi
n

i=1
i=1
! 12
* n
+
n
X
X
a
2
kxi k
Re
xi ,
n
kak
i=1
i=1
* n
+
X
r2
a

Re
q
xi ,
q
kak
2
2
2
2
i=1
kak r kak + kak r
n

X
r2

q
xi .

q

i=1
kak2 r2 kak + kak2 r2
n
X
We note that for kak = 1 and r (0, 1) , the inequality (6.89) becomes
! 12
n
n
X
X
p
(6.115)
1 r2
kxi k (1 r2 ) n
kxi k2
i=1
i=1
Re
* n
X
+
xi , a
i=1
253

n
X

xi

i=1
which is a refinement of (6.107).

With the same assumptions for a and r, we have from (6.114) the
following additive reverse of the generalised triangle inequality:
0
(6.116)
n
X
i=1

n
X

kxi k
xi

i=1
* n
+
X
r2
Re
xi , a
1 + 1 r2
i=1

n

2
X
r

xi .
2

1+ 1r
i=1
1 r2
1 r2
We can obtain the following reverses of the generalised triangle

inequality from Corollary 54 when the assumptions are in terms of
complex numbers and :
If , K with Re (
) > 0 and xi H, i {1, . . . , n} , e H,
kek = 1 are such that
(6.117)

xi + e 1 | | for each i {1, . . . , n} ,

2
2
or, equivalently,
Re he xi , xi ei 0 for each i {1, . . . , n} ,
then we have the following reverses of the generalised triangle inequality:
(6.118)
n
X
kxi k
n
X
i=1
! 12
2
kxi k
i=1
P

+
h ni=1 xi , ei
p
2 Re (
)

n

X
+

1

p
xi

2
Re (
) i=1
Re
254
and
(6.119)

n
X

0
kxi k
xi

i=1
i=1

! 12 n
n
X
X

n
kxi k2

xi

i=1
i=1
+
! 12

* n
n

X
X
Re q
xi , e
kxi k2
n
Re
i=1
i=1
n
X
| |2

p
p
2 Re (
) | + | + 2 Re (
)
"
* n
+#
+
X

Re
xi , e
+
i=1

n
X

p
xi .
p

2 Re (
) | + | + 2 Re (
)
| |2
i=1
Obviously (6.118) for = M, = m, M m > 0 provides a

refinement for (6.111).
6.2.5. Lower Bounds for the Distance to Finite-Dimensional
Subspaces. Let (H; h, i) be an inner product space over the real or
complex number field K, {y1 , . . . , yn } a subset of H and G (y1 , . . . , yn )
the Gram matrix of {y1 , . . . , yn } where (i, j) entry is hyi , yj i . The determinant of G (y1 , . . . , yn ) is called the Gram determinant of {y1 , . . . , yn }
and is denoted by (y1 , . . . , yn ) .
Following [4, p. 129 133], we state here some general results for
the Gram determinant that will be used in the sequel:
(1) Let {x1 , . . . , xn } H. Then (x1 , . . . , xn ) 6= 0 if and only if
{x1 , . . . , xn } is linearly independent;
(2) Let M = span {x1 , . . . , xn } be ndimensional in H, i.e., {x1 , . . . ,
xn } is linearly independent. Then for each x H, the distance
d (x, M ) from x to the linear subspace H has the representations
(6.120)
d2 (x, M ) =
(x1 , . . . , xn , x)
(x1 , . . . , xn )
255
and
(6.121)
P
2
( ni=1 |hx,xi i|2 )
if x
/ M ,
kxk Pn
2
k i=1 hx,xi ixi k
2
d (x, M ) =
kxk2
if x M ,
where M denotes the orthogonal complement of M.

The following result may be stated [11].
Proposition 60. Let {x1 , . . . , xn } be a system of linearly independent vectors, M = span {x1 , . . . , xn } , x H\M , a H, r > 0 and
kak > r. If

hx,
x
ia
(6.122)
x
i
i |hx, xi i| r for each i {1, . . . , n} ,
(note that if hx, xi i 6= 0 for each i {1, . . . , n} , then (6.122) can be
written as

x

i
(6.123)
a r for each i {1, . . . , n}),

hx, xi i
(6.124)
Pn
2
kak2
i=1 |hx, xi i|
d (x, M ) kxk
P
n
2
kak2 r2
i=1 kxi k
0.
2
Proof. Utilising (6.121) we can state that

Pn
n
X
|hx, xi i|2
2
2
(6.125)
d (x, M ) = kxk Pni=1
|hx, xi i|2 .
2
Also, by the inequality (6.86) applied for i = hx, xi i , pi =
{1, . . . , n} , we can state that
Pn
2
1
kak2
i=1 |hx, xi i|
(6.126)
Pn
Pn
2
2
2
2
kak r
k i=1 hx, xi i xi k
i=1 kxi k
1
,
n
provided the condition (6.123) holds true.

Combining (6.125) with (6.126) we deduce the first inequality in
(6.124).
The last inequality is obvious since, by Schwarzs inequality
n
n
n
X
X
kak2 X
2
2
2
kxk
kxi k
|hx, xi i|
|hx, xi i|2 .
2
2
kak r i=1
i=1
i=1
256
Remark 73. Utilising (6.120), we can state the following result for
Gram determinants
(6.127) (x1 , . . . , xn , x)
"
#
Pn
2
2
|hx,
x
i|
kak
i
i=1
P
(x1 , . . . , xn ) 0
kxk2
n
2
kak2 r2
i=1 kxi k
for x
/ M and x, xi , a and r are as in Proposition 60.
The following corollary of Proposition 60 may be stated as well [11].
Corollary 55. Let {x1 , . . . , xn } be a system of linearly independent vectors, M = span {x1 , . . . , xn } , x H\M and , K with
Re (
) > 0. If e H, kek = 1 and

+
1

(6.128)
xi hx, xi i 2 e 2 | | |hx, xi i|
or, equivalently,
E
D
Re hx, xi ie xi , xi hx, xi ie 0,
for each i {1, . . . , n} , then
(6.129)
Pn
2
1 | + |2
i=1 |hx, xi i|
Pn
d (x, M ) kxk
0,
2
4 Re (
)
i=1 kxi k
2
or, equivalently,
(6.130) (x1 , . . . , xn , x)
#
"
2
2 Pn
1 | + |
i=1 |hx, xi i|
kxk2
P
(x1 , . . . , xn ) 0.
n
2
4 Re (
)
i=1 kxi k
6.2.6. Applications for Fourier Coefficients. Let (H; h, i) be
a Hilbert space over the real or complex number field K and {ei }iI an
orthornormal basis for H. Then (see for instance [4, p. 54 61])
(i) Every element x H can be expanded in a Fourier series, i.e.,
X
x=
hx, ei i ei ,
iI
where hx, ei i , i I are the Fourier coefficients of x;

(ii) (Parseval identity)
X
kxk2 =
hx, ei i ei ,
x H;
iI
(iii) (Extended Parseval identity)

X
hx, yi =
hx, ei i hei , yi ,
257
x, y H;
iI
(iv) (Elements are uniquely determined by their Fourier coefficients

)
hx, ei i = hy, ei i for every i I implies that x = y.
Now, we must remark that all the results can be stated for K = K
where K is the Hilbert space of complex (real) numbers endowed with
the usual norm and inner product .
Therefore, we can state the following proposition [11].
Proposition 61. Let (H; h, i) be a Hilbert space over K and
{ei }iI an orthornormal base for H. If x, y H (y 6= 0) , a K (C, R)
and r > 0 such that |a| > r and

hx, ei i

r for each i I,
(6.131)
a
hy, ei i

then we have the following reverse of the Schwarz inequality
1
kxk kyk q
(6.132)
Re [
a hx, yi]
|a|2 r2
|a|
|hx, yi| ;
|a|
(6.133)

a
hx, yi
kxk kyk Re
|a|

a
r2

Re
q
hx, yi
q
|a|
2
2
2
2
|a| r |a| + |a| r
q
(6.134)
r2
r2
2
|a|

r2
kxk2 kyk2
|a| +
|a|
|hx, yi| ;
r2
1
(Re [
a hx, yi])2
2
|a| r
2
|a|2
2
|hx, yi|2
2
|a| r
258
and

2
a
2
2
kxk kyk Re
hx, yi
|a|

2
r2
a
Re
2
hx, yi
|a|
|a| |a|2 r2
(6.135)
r2
|hx, yi| .
|a|2 r2
The proof is similar to the one in Theorem 88, when instead of xi

we take hx, ei i , instead of i we take hei , yi , kk = || , pi = 1, and we
use the Parseval identities mentioned above in (ii) and (iii). We omit
the details.
{ei }iI an orthornormal base for H. If x, y H (y 6= 0) , e, , K
with Re (
) > 0, |e| = 1 and, either

hx, ei i + 1

hy, ei i 2 e 2 | |
(6.136)
or, equivalently,
(6.137)

hx, ei i
hei , xi
Re e
e 0
hy, ei i
hei , yi
for each i I, then the following reverses of the Schwarz inequality

hold:

+
Re
e hx, yi
1
| + |
p
p
|hx, yi| ,
(6.138) kxk kyk
2
2 Re (
)
Re (
)
(6.139)

"
#

+
e
kxk kyk Re
hx, yi
| + |
6.3. OTHER REVERSES OF THE CBS INEQUALITY
259
| |2

p
p
2 Re (
) | + | + 2 Re (
)
#
"

+
e
Re
hx, yi
| + |
| |2

|hx, yi|
p
p
2 Re (
) | + | + 2 Re (
)
and
(6.140)

#)2
( "

+
e
2
2
kxk kyk Re
hx, yi
| + |

2
| |2
+
Re
e hx, yi
2
4 | + | Re (
)
| |2
|hx, yi|2 .
4 Re (
)
Remark 74. If = M m = > 0, then one may state simpler

inequalities from (6.138) (6.140). We omit the details.
6.3. Other Reverses of the CBS Inequality
over the real or complex number field K.
The following reverses for the Schwarz inequality hold (see [8], or
the monograph on line [7, p. 27]).
space over the real or complex number field K. If x, a H and r > 0
are such that
(6.141)
x B (x, r) := {z H| kz ak r} ,
then we have the inequalities

(6.142)
(0 ) kxk kak |hx, ai| kxk kak |Re hx, ai|

1
kxk kak Re hx, ai r2 .
2
The constant 12 is best possible in (6.141) in the sense that it cannot be

replaced by a smaller quantity.
260
An additive version for the Schwarz inequality that may be more

useful in applications is incorporated in [8] (see also [7, p. 28]).
space over K and x, y H and , K with 6= and either
Re hy x, x yi 0,
(6.143)
or, equivalently,
x
| | kyk
y

2
2
(6.144)
holds. Then we have the inequalities

0 kxk kyk |hx, yi|

+

kxk kyk Re
hx, yi
| + |

+
kxk kyk Re
hx, yi
| + |
(6.145)
The constant
1
4
1 | |2
kyk2 .
4 | + |
in the last inequality is best possible.
We remark that a simpler version of the above result may be stated

if one assumed that the scalars are real:
Corollary 56. If M m > 0, and either
Re hM y x, x myi 0,
(6.146)
or, equivalently,

1
m
+
M
x
y

2 (M m) kyk
2
(6.147)
holds, then
0 kxk kyk |hx, yi|
(6.148)
kxk kyk |Re hx, yi|

kxk kyk Re hx, yi
The constant
1
4
is sharp.
1 (M m)2
kyk2 .
4
M +m
261
a Hilbert space over K, pi 0, i N with

PNow, let (K, h, i) be
2
p
=
1.
Consider
`
(K)
as the space
p
i=1 i
(
)
X
`2p (K) := x = (xi ) |xi K, i N and
pi kxi k2 < .
i=1
It is well known that
`2p
(K) endowed with the inner product

hx, yip :=
pi hxi , yi i
i=1
is a Hilbert space over K. The norm kkp of `2p (K) is given by

kxkp :=
! 12
pi kxi k2
i=1
If x, y `2p (K) , then the following Cauchy-Bunyakovsky-Schwarz (CBS)

inequality holds true:

2
X

X
X

(6.149)
pi kxi k2
pi kyi k2
pi hxi , yi i

i=1
i=1
i=1
with equality iff there exists a K such that xi = yi for each i N.

If
(
)
X

`2p (K) := = (i ) i K, i N and
pi |i |2 <
iN
i=1
and x
(6.150)
`2p
(K) , then the following (CBS)-type inequality is also valid:

2
X

X
X

2
2
pi |i |
pi kxi k
pi i xi

i=1
i=1
i=1
with equality if and only if there exists a vector v K such that

xi = i v for each i N.
In [11], by the use of some preliminary results obtained in [9],
various reverses for the (CBS)-type inequalities (6.149) and (6.150) for
sequences of vectors in Hilbert spaces were obtained. Applications for
bounding the distance to a finite-dimensional subspace and in reversing
the generalised triangle inequality have also been provided.
The aim of the present section is to provide different results by
employing some inequalities discovered in [8]. Similar applications are
pointed out.
262
6.3.2. Reverses of the (CBS)-Inequality for Two Sequences

in `2p (K). The following proposition may be stated [12].
Proposition 63. Let x, y `2p (K) and r > 0. If
kxi yi k r for each i N,
(6.151)
then
(6.152)
! 21
X

pi kyi k

pi hxi , yi i
(0 )
pi kxi k

i=1
i=1
i=1

! 12
X

X
X

2
2
pi kxi k
pi kyi k

pi Re hxi , yi i

i=1
i=1
i=1
! 12
X
X
X
2
2
pi Re hxi , yi i
pi kxi k
pi kyi k
i=1
i=1
i=1
1
r2 .
2
The constant 12 in front of r2 is best possible in the sense that it cannot
be replaced by a smaller quantity.
Proof. If (6.151) holds true, then
kx
yk2p
X
i=1
pi kxi yi k r
pi = r 2
i=1
and thus kx ykp r.

`2p
Applying the inequality (6.142) for the inner product

(K) , h, ip ,
we deduce the desired result (6.152).
The sharpness of the constant follows by Theorem 89 and we omit
the details.
Proposition 64. Let x, y `2p (K) and , K with 6= . If
either
(6.153)
Re hyi xi , xi yi i 0 for each i N
or, equivalently,

+

1
(6.154)
xi 2 yi 2 | | kyi k for each i N
263
holds, then:
(0 )
(6.155)
pi kxi k
i=1
! 12
pi kyi k
i=1
i=1
pi kxi k
X

pi hxi , yi i

! 12
pi kyi k2
i=1
"i=1
#

+ X

Re
pi hxi , yi i

| + | i=1
! 12
X
X
pi kxi k2
pi kyi k2
i=1
i=1
"
+ X
Re
pi hxi , yi i
| + | i=1
1 | |2 X
pi kyi k2 .

4 | + | i=1
The constant
1
4
Proof. Since, by (6.153),

Re hy x, x yip =
pi Re hyi xi , xi yi i 0,
i=1
hence,
on applying
the inequality (6.145) for the Hilbert space

`2p (K) , h, ip , we deduce the desired inequality (6.155).
The best constant follows by Theorem 90 and we omit the details.
Corollary 57. If the conditions (6.153) and (6.154) hold for =
M, = m with M m > 0, then
(6.156)
! 21

X

pi hxi , yi i
(0 )
pi kxi k
pi kyi k2

i=1
i=1
i=1

! 12
X

X
X

pi kxi k2
pi kyi k2

pi Re hxi , yi i

i=1
X
2
i=1
i=1
264
pi kxi k
i=1
The constant
1
4
X
2
! 21
pi kyi k
i=1
2 X
1 (M m)
4
M +m
pi Re hxi , yi i
i=1
pi kyi k2 .
i=1
is best possible.
6.3.3. Reverses of the (CBS)-Inequality for Mixed Sequences.

Theorem 91 (Dragomir, 2005). Let `2p (K) , x `2p (K) and
v K\ {0} , r > 0. If
kxi i vk r |i | for each i N
(6.157)
(note that if i 6= 0 for any i N, then the condition (6.157) is equivalent to the simpler one

xi

(6.158)
i v r for each i N),
then
(6.159)
! 12
X

(0 )

pi i xi

i=1
i=1
i=1
! 12 *
+
X

X
X
v

pi |i |2
pi kxi k2

pi i xi ,

kvk
i=1
i=1
i=1
! 21 *
+

X
X
X
v

2
2
pi |i |
pi kxi k
Re
pi i xi ,

kvk
i=1
i=1
i=1
*
! 21
+
X
X
X
v
2
2
Re
pi i xi ,
pi |i |
pi kxi k
kvk
i=1
i=1
i=1
X
pi |i |2
pi kxi k2
The constant
1
2
1 r2 X
pi |i |2 .
2 kvk i=1

kxi k2 2 Re hi xi , vi + |i |2 kvk2 r2 |i |2 ,
(6.160)
kxi k2 + |i |2 kvk2 2 Re hi xi , vi + r2 |i |2
265
for each i N.
If we multiply (6.160) by pi 0, i N and sum over i N, then
we deduce
(6.161)
pi kxi k2 + kvk2
i=1
pi |i |2
i=1
2 Re
*
X
+
pi i xi , v
+r
i=1
pi |i |2 .
i=1
Since, obviously
(6.162)
2 kvk
X
2
pi |i |
pi kxi k2
i=1
i=1
! 12
pi kxi k2 + kvk2
i=1
pi |i |2 ,
i=1
hence, by (6.161) and (6.162), we deduce

! 12
X
X
2 kvk
pi |i |2
pi kxi k2
i=1
i=1
2 Re
*
X
+
pi i xi , v
i=1
+r
pi |i |2 ,
i=1
which is clearly equivalent to the last inequality in (6.159).

The other inequalities are obvious.
The best constant follows by Theorem 89.
The following corollary may be stated [12].
Corollary 58. Let `2p (K) , x `2p (K) , e H, kek = 1 and
, K with 6= . If

1
xi i
(6.163)
e

2 | | |i |
2
for each i N, or, equivalently,
(6.164)
Re hi e xi , xi i ei
for each i N (note that, if i 6= 0 for any i N, then (6.163) is

equivalent to

xi
+

1
(6.165)
i 2 e 2 | |
266
for each i N and (6.164) is equivalent to

xi xi
(6.166)
Re e ,
e 0
i i
for each i N), then the following reverse of the (CBS)-inequality is
valid:

! 12
X

X
X

(6.167)

pi i xi
(0 )
pi |i |2
pi kxi k2

i=1
i=1
i=1
! 12 *
+
X

X
X

pi |i |2
pi kxi k2

pi i xi , e

i=1
i=1
i=1
! 12
X
X
pi |i |2
pi kxi k2
i=1
i=1
"
*
+#

X

Re
pi i xi , e

| + | i=1
! 12
X
X
pi |i |2
pi kxi k2
i=1
i=1
"
Re
| + |
The constant
1
4
*
X
+#
pi i xi , e
i=1
1 | |2 X
pi |i |2 .
4 | + | i=1
is best possible.
Remark 75. If M m > 0, i 6= 0 and for e as above, either

xi
1
M
+
m

(6.168)
e
i
2 (M m) for each i N
2
or, equivalently,

xi xi
Re M e ,
me 0 for each i N
i i
holds, then
(0 )
X
2
pi |i |
pi kxi k2
i=1
i=1
! 12
X

pi i xi

i=1
! 12
*
+
X

pi i xi , e

i=1
i=1
i=1
+
! 12 *

X
X
X

2
2
Re
pi |i |
pi kxi k
pi i xi , e

i=1
i=1
i=1
! 21
*
+
X
X
X
2
2
Re
pi i xi , e
pi |i |
pi kxi k
267
X
2
pi |i |
pi kxi k2
i=1
i=1
2 X
1 (M m)
4
M +m
The constant
1
4
i=1
pi |i |2 .
i=1
is best possible.
6.3.4. Reverses for the Generalised Triangle Inequality. In

1966, Diaz and Metcalf [5] proved the following interesting reverse of
the generalised triangle inequality:

X
X

(6.169)
r
kxi k
xi ,

i=1
i=1
provided the vectors x1 , . . . , xn H\ {0} satisfy the assumption

Re hxi , ai
(6.170)
0r
,
i {1, . . . , n} ,
kxi k
where a H, kak = 1 and (H; h, i) is a real or complex inner product
space.
In an attempt to provide other sufficient conditions for (6.169) to
hold, the author pointed out in [14] that

X

X
p

(6.171)
1 2
kxi k
xi

i=1
i=1
where the vectors xi , i {1, . . . , n} satisfy the condition

(6.172)
kxi ak ,
i {1, . . . , n} ,
where r H, kak = 1 and (0, 1) .

Following [14], if M m > 0 and the vectors xi H, i {1, . . . , n}
verify either
(6.173)
Re hM a xi , xi mai 0,
or, equivalently,

M +m
1

(6.174)
a
(M m) ,
xi

2
2
i {1, . . . , n} ,
i {1, . . . , n} ,
268
where a H, kak = 1, then

n
n
X

2 mM X

kxi k
(6.175)
xi .

M + m i=1
i=1
It is obvious from Theorem 91, that, if
(6.176)
kxi vk r,
i {1, . . . , n} ,
for
where xi H, i {1, . . . , n} , v H\ {0} and r > 0, then we can state

the inequality

! 12 n
n
1 X
1X

(0 )
(6.177)
kxi k2

xi
n

n i=1
i=1
! 12 * n
+
n

1X
v
1X

kxi k2

xi ,

n
n i=1
kvk
i=1
! 12 * n
+
n

X
1X
1
v

2
kxi k
Re
xi ,

n i=1
n i=1
kvk
! 12
* n
+
n
X
X
1
1
v
kxi k2
Re
xi ,
n i=1
n i=1
kvk
1 r2
.
2 kvk
Since, by the (CBS)-inequality we have

n
(6.178)
1X
kxi k
n i=1
1X
kxi k2
n i=1
! 12
,
hence, by (6.177) and (6.173) we have [12]:

n
n
X
1
X
r2

(6.179)
(0 )
kxi k
xi n

2
kvk
i=1
i=1
provided that (6.176) holds true.
Utilising Corollary 58, we may state that, if

+

1
(6.180)
i {1, . . . , n} ,
xi 2 e 2 | | ,
or, equivalently,
(6.181)
Re he xi , xi ei 0,
i {1, . . . , n} ,
269
where e H, kek = 1, , K, 6= and xi H, i {1, . . . , n} ,

then

! 12 n
n
1 X
1X

2
(6.182)

xi
(0 )
kxi k

n
n i=1
i=1
+
! 12 * n
n
1X

1X

xi , e
kxi k2

n
n i=1
i=1
! 12 "
* n
+#
n

X
1
1X

kxi k2
Re
xi , e

n i=1
| + | n i=1
! 12
"
* n
+#
n
X
1
1X
kxi k2
Re
xi , e
n i=1
| + | n i=1
1 | |2

.
4 | + |
Now, making use of (6.178) and (6.182) we can establish the following
additive reverse of the generalised triangle inequality [12]

n
n
X
1
X
| |2

kxi k
xi n
(6.183)
(0 )
,

4
| + |
i=1
i=1
provided either (6.180) or (6.181) hold true.
6.3.5. Applications for Fourier Coefficients. Let (H; h, i) be
a Hilbert space over the real or complex number field K and {ei }iI an
orthonormal basis for H. Then (see for instance [4, p. 54 61]):
(i) Every element x H can be expanded in a Fourier series, i.e.,
X
x=
hx, ei i ei ,
iI
where hx, ei i , i I are the Fourier coefficients of x;

(ii) (Parseval identity)
X
kxk2 =
hx, ei i ei , x H;
iI
(iii) (Extended Parsevals identity)

X
hx, yi =
hx, ei i hei , yi ,
iI
x, y H;
270
(iv) (Elements are uniquely determined by their Fourier coefficients)

hx, ei i = hy, ei i
for every i I implies that x = y.
We must remark that all the results from the second and third sections may be stated for K = K where K is the Hilbert space of complex
(real) numbers endowed with the usual norm and inner product .
Therefore we can state the following reverses of the Schwarz inequality [12]:
{ei }iI an orthonormal base for H. If x, y H, y 6= 0, a K (C, R)
with r > 0 such that

hx, ei i

for each i I,
(6.184)
hy, ei i a r
then we have the following reverse of the Schwarz inequality:

a

kxk kyk Re hx, yi
|a|

a
kxk kyk Re hx, yi

|a|
2
1 r

kyk2 .
2 |a|
(6.185)
The constant
1
2
The proof is similar to the one in Theorem 91, where instead of xi

we take hx, ei i, instead of i we take hei , yi , kk = || , pi = 1 and use
the Parseval identities mentioned above in (ii) and (iii). We omit the
details.
{ei }iI an orthonormal base for H. If x, y H, y 6= 0, e, , K
with |e| = 1, 6= and

hx, ei i + 1

(6.186)
hy, ei i 2 e 2 | |
or equivalently,
(6.187)

hx, ei i
hei , xi
Re e
e 0
hy, ei i
hei , yi
271
for each i I, then


+

kxk kyk Re
hx, yi e
| + |

+
kxk kyk Re
hx, yi e
| + |
(6.188)
1 | |2

kyk2 .
4 | + |
The constant
1
4
is best possible.
Remark 76. If = M m = > 0, then one may state simpler

inequalities from (6.188). We omit the details.
Bibliography
[1] R. BELLMAN, Almost orthogonal series, Bull. Amer. Math. Soc., 50 (1944),
517-519.
[2] R.P. BOAS, A general moment problem, Amer. J. Math., 63 (1941), 361-370.
[3] E. BOMBIERI, A note on the large sieve, Acta Arith., 18 (1971), 401-404.
[4] F. DEUTSCH, Best Approximation in Inner Product Spaces, CMS Books in
Mathematics, Springer-Verlag, New York, Berlin, Heidelberg, 2001.
Hilbert and Banach spaces, Proc. Amer. Math. Soc., 17(1) (1966), 88-99.
[6] S.S. DRAGOMIR, A counterpart of Bessels inequality in inner product spaces
and some Gr
uss type related results, RGMIA Res. Rep. Coll., 6 (2003), Supplement, Article 10. [ONLINE: http://rgmia.vu.edu.au/v6(E).html].
[7] S.S. DRAGOMIR, Advances in Inequalities of the Schwarz, Gr
uss and Bessel
Type in Inner Product Spaces, RGMIA Monographs, Victoria University, 2004.
[ONLINE http://rgmia.vu.edu.au/monographs/advancees.htm].
[8] S.S. DRAGOMIR, New reverses of Schwarz, triangle and Bessel inequalities in
inner product spaces, Australian J. Math. Anal. & Appl., 1(1) (2004), Art. 1.
[ONLINE http://ajmaa.org/].
[9] S.S. DRAGOMIR, Reverses of Schwarz, triangle and Bessel inequalities in inner
product spaces, J. Ineq. Pure & Appl. Math., 5(3) (2004), Art. 74. [ONLINE
http://jipam.vu.edu.au/article.php?sid=432].
spaces, RGMIA Res. Rep. Coll., 7 (2004), Supplement, Article 7. [ONLINE:
[11] S.S. DRAGOMIR, Reversing the CBS-inequality for sequences of vectors in
Hilbert spaces with applications (I), RGMIA Res. Rep. Coll., 8(2005), Supplement, Article 2. [ONLINE http://rgmia.vu.edu.au/v8(E).html].
[12] S.S. DRAGOMIR, Reversing the CBS-inequality for sequences of vectors in
Hilbert spaces with applications (II), RGMIA Res. Rep. Coll., 8(2005), Supplement, Article 3. [ONLINE http://rgmia.vu.edu.au/v8(E).html].
[13] S.S. DRAGOMIR, Some Bombieri type inequalities in inner product spaces,
J. Indones. Math. Soc., 10(2) (2004), 91-97.
spaces, RGMIA Res. Rep. Coll., 7 (2004), Supplement, Article 7. [ONLINE
[15] S.S. DRAGOMIR, On the Boas-Bellman inequality in inner product spaces,
Bull. Austral. Math. Soc., 69(2) (2004), 217-225.
[16] S.S. DRAGOMIR, Upper bounds for the distance to finite-dimensional subspaces in inner product spaces, RGMIA Res. Rep. Coll., 7(1) (2005), Article
2. [ONLINE: http://rgmia.vu.edu.au/v7n1.html].
273
274
BIBLIOGRAPHY
J.E. PECARI
and A.M. FINK, Classical and New In[17] D.S. MITRINOVIC,

C
equalities in Analysis, Kluwer Academic, Dordrecht, 1993.
Index
absolutely continuous, 86, 210
Dragomir-Mond, 9, 11, 1318, 24, 26,

27, 30, 3234
Dragomir-Sandor, 17, 21, 22, 41
Banach space, 151

Bellman, 231
Bessels inequality, 1618, 20, 45, 59,
61, 63, 71, 76, 139, 157, 183, 226,
231, 233, 235, 236, 238
binary relation, 8, 9
Blatter, 55
Boas, 231
Bochner integrable, vi, 86, 152, 210
Bochner measurable, 83, 142, 152, 197
Bombieri, 235
Buzano, 5155, 58, 61, 67
field, 1, 21, 27, 37, 43, 53, 59, 66, 74,

7880, 83, 88, 89, 92, 97, 107,
112, 114, 116, 122, 129, 133, 152,
166, 197, 203, 216, 225, 236, 239,
251, 254, 256, 259, 269
Fourier coefficients, 256, 257, 269, 270
Fujii, 52
functional, 13, 513, 16, 19, 20, 24,
27, 28, 30, 33, 37, 41
generalised triangle inequality, 89, 101,
102, 107, 108, 112, 116, 124, 125,
130, 133, 134, 138, 139, 142, 143,
145, 151, 241, 251, 253, 261, 267,
269
Goldstein, 91
Gram determinants, v, 1, 21, 225, 226,
254, 256
Gram matrix, 21, 225, 254
Grams inequality, 21, 226
Clarke, 91
complex function, 56
complex numbers, vi, 32, 57, 107, 108,
145, 240, 253
complex sequence, 58
complexification, 6, 7, 38, 46, 50, 53,
56, 57, 63, 73
convex cone, 8
de Bruijn, 46, 52, 56
Diaz-Metcalf, 107, 108, 111, 138
discrete inequality, 52, 56, 80
Dragomir, 3840, 43, 48, 50, 5355,
57, 59, 6163, 68, 71, 72, 74, 76,
78, 79, 86, 88, 94, 96, 97, 99, 101,
108, 109, 112, 114, 116119, 122,
128, 129, 131, 133135, 138, 153,
156, 160, 166, 172, 173, 175, 176,
178, 183, 198, 199, 204, 206, 210
213, 215, 217, 219, 221, 222, 227,
228, 230, 232237, 239, 240, 245,
259, 260, 264
Hadamards inequality, 21, 22, 226, 228,

232, 233, 235, 236
Heisenberg, 38, 86, 87, 198, 210, 211,
213, 221, 222
Heisenberg inequality, v, vi, 38, 87,
198, 211, 213, 222
Hermitian form, 1, 2, 12, 16, 38
Hilbert space, v, vi, 1, 25, 38, 41, 56,
58, 64, 76, 8084, 86, 142144,
152, 154, 166, 178181, 183, 185,
189, 193, 197, 198, 204, 210, 217,
275
276
218, 241, 244, 256258, 261, 263,

269, 270
Hile, 91
index set, 1620, 28, 33
inner product, 1, 15, 21, 23, 27, 28,
37, 38, 41, 43, 46, 4851, 53, 55
57, 59, 61, 63, 6568, 72, 74, 78
80, 83, 88, 89, 9299, 101, 103,
107109, 112, 114, 116119, 122,
125, 127129, 131, 133138, 141,
142, 146, 152, 153, 193, 197, 199,
203, 204, 214, 216, 217, 225, 226,
231, 235237, 239, 241, 242, 251,
254, 257, 259262, 267, 270
Karamata, vi, 107, 151, 187, 192
Kroneckers delta, 15, 76, 226
Kubo, 52
Kurepa, v, 2, 5, 6, 38, 46, 47, 50, 51,
53, 56, 57, 59, 63, 64, 72
Kurepas inequality, 74
linear combination, 7
linear space, 1, 5, 6, 8, 19, 27, 37, 47
linear subspace, 2, 4, 41, 225, 254
linearly dependent, 21, 37, 226
linearly independent, vi, 7, 225228,
237, 254256
lower bounds, 37, 38
Marden, 107, 151
Metcalf, 107, 151, 251, 267
modulus, 39, 54, 55, 71
monotonicity, 1, 14, 2426
Moore, v, 38, 6668
n-dimensional, 225, 254
nondecreasing, 913, 18, 19, 2729,
277
norm, 24, 37, 41, 46, 80, 81, 83, 84,
142, 197, 241, 257, 261, 270
order, 9
orthogonal, 21, 38, 41, 45, 52, 114,
139, 226, 228, 255
orthonormal base, 270
orthonormal family, 15, 43, 46, 59,
6163, 71, 73, 76, 138, 166, 168
171, 238
INDEX
Parseval, 256258, 269, 270

Petrovich, 107, 151
positive definite, 1, 25, 26, 57
positive semi-definite, 13, 5, 8
Precupanu, v, 38, 55, 59, 6567, 70,
72
quadratic reverses, 122, 125, 172, 191
real function, 56, 58, 191
real numbers, 5, 10, 40, 46, 56, 58, 98
real sequence, 27
refinement, 1721, 23, 26, 31, 34, 37
40, 42, 43, 46, 61, 74, 92, 231
233, 235, 253, 254
Bessels inequality, 76
Buzano inequality, 53
Buzanos inequality, 55, 61
CBS inequality, 52, 56, 80, 83
CBS integral inequality, 85
Heisenberg inequality, 87
Kurepas inequality, 64
Kurepas result, 57
Schwarz inequality, 78, 79, 81, 82,
173
quadratic, 93
Schwarzs inequality, 56, 67, 74, 78,
123
triangle inequality, 78, 124, 173
reverse, 37, 39, 66, 8890, 9294, 97,
99102, 107, 108, 111, 112, 116,
121, 122, 125, 130, 132134, 136,
138, 139, 142, 143, 145, 151, 152,
159, 166, 172, 180, 186, 189192,
204207, 211, 213, 215, 217, 221,
222, 236, 238241, 245, 248, 250,
251, 253, 257259, 261, 266, 267,
269, 270, 273
Richards, 38
Ryff, 91
scalar product, 46, 47
Schwarz inequality, v, 1, 10, 18, 20,
23, 26, 3740, 42, 43, 45, 49, 52,
56, 65, 66, 70, 7882, 8890, 92,
94, 97, 99101, 121, 122, 131, 135,
153, 173, 179, 230, 239, 242, 257
260, 270
quadratic, 93
INDEX
self-adjoint operator, 25
strong nondecreasing, 11, 18
strong superadditive, 11, 17
superadditivity, 1, 14, 15, 18, 20, 23,
26
supremum, 40, 41, 51
triangle inequality, v, vi, 10, 37, 55,
64, 71, 78, 89, 101, 102, 107, 108,
111, 112, 116, 121, 122, 124, 125,
130, 132134, 138, 139, 142, 143,
145, 151, 152, 159, 166, 172, 173,
186, 189192, 241, 251, 253, 261,
267, 269
upper bounds, vi, 37, 88, 89, 108, 226
vector-valued function, vi, 198, 204,
217
vectors, 1, 38, 1517, 21, 23, 27, 28,
37, 38, 4143, 45, 47, 59, 6668,
72, 88, 89, 101, 103, 107109, 111,
114, 116, 118, 120, 122, 135139,
141, 152157, 159161, 163, 166,
180, 181, 183, 185, 195, 198, 204,
206, 217, 225228, 231, 235239,
241, 244, 251, 255, 256, 261, 267
Wilf, 107, 151
277

Sever S. Dragomir-Advances in Inequalities of The Schwarz

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

Sever S. Dragomir-Advances in Inequalities of The Schwarz

Transféré par

Droits d'auteur :

Formats disponibles

Advances in Inequalities of the

Schwarz, Triangle and Heisenberg

2000 Mathematics Subject Classification. Primary 46C05, 46E30;

Chapter 1. Inequalities for Hermitian Forms

Chapter 2. Schwarz Related Inequalities

Chapter 3. Reverses for the Triangle Inequality

Chapter 4. Reverses for the Continuous Triangle Inequality

Chapter 5. Reverses of the CBS and Heisenberg Inequalities

Chapter 6. Other Inequalities in Inner Product Spaces

related to the Diaz-Metcalf results, or in the equivalent quadratic form

Inequalities for Hermitian Forms

1. INEQUALITIES FOR HERMITIAN FORMS

1.2. Hermitian Forms, Fundamental Properties

and L (z) := {az|a K} .

The following fundamental facts concerning Hermitian forms hold

|(x, y)|2 (x, x) (y, y)

holds for all x K and all y X;

|(x, y)|2 (x, x) (y, y)

holds for all x, y X;

1.2. HERMITIAN FORMS, FUNDAMENTAL PROPERTIES

|(e, y)|2 (e, e) (y, y)

satisfies the conditions

(x, x) = |a|2 (e0 , e0 ) + (x0 , x0 ) 0.

Therefore (, ) is a positive semi-definite functional on X (f ) .

1. INEQUALITIES FOR HERMITIAN FORMS

From the parallelogram identity:

we conclude that the set X0 = X0 X (e) is a linear subspace of X.

where X+ is any direct complement of X0 in the space X (e) . If y 6= 0,

is in X (e) and therefore (y 0 , y 0 ) 0.

Hence y X+ implies that (e0 , y) = 0, i.e.,

which together with y X (e) leads to (f, y) = 0. Thus y X+ implies

1.2. HERMITIAN FORMS, FUNDAMENTAL PROPERTIES

[Re (e, y)]2 (e, e) (y, y) ,

[Re (e, y)]2 (e, e) (y, y) ,

holds for all y X and for an e X.

[Re (e, y)]2 (e, e) (y, y) = 2 (e, e)2 (e, e) (x, x) .

If (, ) is semi-definite on X, then (1.11) implies (1.9).

1. INEQUALITIES FOR HERMITIAN FORMS

1.2.2. Schwarzs Inequality for the Complexification of a

Formally, we can write z = x + iy = Re z + i Im z and z = x iy =

for all y X. Then for x, y X we have

|(e + if, x + iy)C |2 (e + if, e + if )C (x + iy, x + iy)C

1.2. HERMITIAN FORMS, FUNDAMENTAL PROPERTIES

holds for all x, y X and that

(e + if, e + if )C = (e, e) + (f, f ) < 0.

If e = af with a real number a, then (1.13) implies that (f, f ) < 0

and (1.13) implies

D (e, e) + D (f, f ) < 0.

Further, consider in Y a base consisting of the two vectors {u1 , u2 } on

belongs to X + . From this it follows that

1. INEQUALITIES FOR HERMITIAN FORMS

Now, replacing in (1.12) the vector x + iy with z X + , we get from

If (, ) H (X) , then the functional kk = (, ) 2 is a semi-norm on

kxk2 kyk2 |(x, y)|2

kxk kyk |(x, y)|

We observe that the following properties hold:

1.3. SUPERADDITIVITY AND MONOTONICITY

i.e., the binary relation defined by (1.18) is an order relation on

1.3.2. The Superadditivity and Monotonicity of Mapping.

((, ) ; x, y) := kxk kyk |(x, y)| ,

which is closely related to Schwarzs inequality (1.17).

for all x, y X, i.e., the mapping (; x, y) is superadditive

for all x, y X, i.e., the mapping (; x, y) is nondecreasing