Skip to main content

Norm inequalities of Čebyšev type for power series in Banach algebras

Abstract

Let f(λ)= n = 0 α n λ n be a function defined by power series with complex coefficients and convergent on the open disk D(0,R)C, R>0 and x,yB, a Banach algebra, with xy=yx. In this paper we establish some upper bounds for the norm of the Čebyšev type difference f(λ)f(λxy)f(λx)f(λy), provided that the complex number λ and the vectors x,yB are such that the series in the above expression are convergent. Applications for some fundamental functions such as the exponential function and the resolvent function are provided as well.

MSC: 47A63, 47A99.

1 Introduction

For two Lebesgue integrable functions f,g:[a,b]R, consider the Čebyšev functional:

C(f,g):= 1 b a a b f(t)g(t)dt 1 ( b a ) 2 a b f(t)dt a b g(t)dt.
(1.1)

In 1935, Grüss [1] showed that

|C(f,g)| 1 4 (Mm)(Nn),
(1.2)

provided that there exist real numbers m, M, n, N such that

mf(t)Mandng(t)Nfor a.e. t[a,b].
(1.3)

The constant 1 4 is best possible in (1.1) in the sense that it cannot be replaced by a smaller quantity.

Another, however, less known result, even though it was obtained by Čebyšev in 1882 [2], states that

|C(f,g)| 1 12 f g ( b a ) 2 ,
(1.4)

provided that f , g exist and are continuous on [a,b] and f = sup t [ a , b ] | f (t)|. The constant 1 12 cannot be improved in the general case.

The Čebyšev inequality (1.4) also holds if f,g:[a,b]R are assumed to be absolutely continuous and f , g L [a,b], while f =ess sup t [ a , b ] | f (t)|.

A mixture between Grüss’ result (1.2) and Čebyšev’s one (1.4) is the following inequality obtained by Ostrowski in 1970 [3]:

|C(f,g)| 1 8 (ba)(Mm) g ,
(1.5)

provided that f is Lebesgue integrable and satisfies (1.3), while g is absolutely continuous and g L [a,b]. The constant 1 8 is best possible in (1.5).

The case of Euclidean norms of the derivative was considered by Lupaş in [4] in which he proved that

|C(f,g)| 1 π 2 f 2 g 2 (ba),
(1.6)

provided that f, g are absolutely continuous and f , g L 2 [a,b]. The constant 1 π 2 is the best possible.

Recently, Cerone and Dragomir [5] have proved the following results:

|C(f,g)| inf γ R g γ q 1 b a ( a b | f ( t ) 1 b a a b f ( s ) d s | p d t ) 1 p ,
(1.7)

where p>1 and 1 p + 1 q =1 or p=1 and q=, and

|C(f,g)| inf γ R g γ 1 1 b a ess sup t [ a , b ] |f(t) 1 b a a b f(s)ds|,
(1.8)

provided that f L p [a,b] and g L q [a,b] (p>1, 1 p + 1 q =1; p=1, q= or p=, q=1).

Notice that for q=, p=1 in (1.7) we obtain

| C ( f , g ) | inf γ R g γ 1 b a a b | f ( t ) 1 b a a b f ( s ) d s | d t g 1 b a a b | f ( t ) 1 b a a b f ( s ) d s | d t
(1.9)

and if g satisfies (1.3), then

| C ( f , g ) | inf γ R g γ 1 b a a b | f ( t ) 1 b a a b f ( s ) d s | d t g n + N 2 1 b a a b | f ( t ) 1 b a a b f ( s ) d s | d t 1 2 ( N n ) 1 b a a b | f ( t ) 1 b a a b f ( s ) d s | d t .
(1.10)

The inequality between the first and the last term in (1.10) has been obtained by Cheng and Sun in [6]. However, the sharpness of the constant 1 2 , a generalization for the abstract Lebesgue integral, and the discrete version of it have been obtained in [7].

For other recent results on the Grüss inequality, see [822], and the references therein.

In order to consider a Čebyšev type functional for functions of vectors in Banach algebras, we need some preliminary definitions and results as follows.

2 Some facts on Banach algebras

Let be an algebra. An algebra norm on is a map :B[0,) such that (B,) is a normed space, and, further

abab

for any a,bB. The normed algebra (B,) is a Banach algebra if is a complete norm.

We assume that the Banach algebra is unital, this means that has an identity 1 and that 1=1.

Let be a unital algebra. An element aB is invertible if there exists an element bB with ab=ba=1. The element b is unique; it is called the inverse of a and written a 1 or 1 a . The set of invertible elements of is denoted by InvB. If a,bInvB then abInvB and ( a b ) 1 = b 1 a 1 .

For a unital Banach algebra we also have:

  1. (i)

    if aB and lim n a n 1 / n <1, then 1aInvB;

  2. (ii)

    {aB:1b<1}InvB;

  3. (iii)

    InvB is an open subset of ;

  4. (iv)

    the map InvBa a 1 InvB is continuous.

For simplicity, we denote λ 1, where λC and 1 is the identity of , by λ. The resolvent set of aB is defined by

ρ(a):={λC:λaInvB};

the spectrum of a is σ(a), the complement of ρ(a) in , and the resolvent function of a is R a :ρ(a)InvB, R a (λ):= ( λ a ) 1 . For each λ,γρ(a) we have the identity

R a (γ) R a (λ)=(λγ) R a (λ) R a (γ).

We also have σ(a){λC:|λ|a}. The spectral radius of a is defined as ν(a)=sup{|λ|:λσ(a)}.

If a, b are commuting elements in , i.e. ab=ba, then

ν(ab)ν(a)ν(b)andν(a+b)ν(a)+ν(b).

Let a unital Banach algebra and aB. Then

  1. (i)

    the resolvent set ρ(a) is open in ;

  2. (ii)

    for any bounded linear functionals λ: B C, the function λ R a is analytic on ρ(a);

  3. (iii)

    the spectrum σ(a) is compact and nonempty in ;

  4. (iv)

    for each nN and r>ν(a), we have

    a n = 1 2 π i | ξ | = r ξ n ( ξ a ) 1 dξ;
  5. (v)

    we have ν(a)= lim n a n 1 / n .

Let f be an analytic functions on the open disk D(0,R) given by the power series f(λ):= j = 0 α j λ j (|λ|<R). If ν(a)<R, then the series j = 0 α j a j converges in the Banach algebra because j = 0 | α j | a j <, and we can define f(a) to be its sum. Clearly f(a) is well defined and there are many examples of important functions on a Banach algebra that can be constructed in this way. For instance, the exponential map on denoted exp and defined as

expa:= j = 0 1 j ! a j for each aB.

If is not commutative, then many of the familiar properties of the exponential function from the scalar case do not hold. The following key formula is valid, however, with the additional hypothesis of commutativity for a and b from :

exp(a+b)=exp(a)exp(b).

In a general Banach algebra it is difficult to determine the elements in the range of the exponential map exp(B), i.e. the element which have a ‘logarithm’. However, it is easy to see that if a is an element in B such that 1a<1, then a is in exp(B). That follows from the fact that if we set

b= n = 1 1 n ( 1 a ) n ,

then the series converges absolutely and, as in the scalar case, substituting this series into the series expansion for exp(b) yields exp(b)=a.

It is well known that if x and y are commuting, i.e. xy=yx, then the exponential function satisfies the property

exp(x)exp(y)=exp(y)exp(x)=exp(x+y).

Also, if x is invertible and a,bR with a<b then

a b exp(tx)dt= x 1 [ exp ( b x ) exp ( a x ) ] .

Moreover, if x and y are commuting and yx is invertible, then

0 1 exp ( ( 1 s ) x + s y ) d s = 0 1 exp ( s ( y x ) ) exp ( x ) d s = ( 0 1 exp ( s ( y x ) ) d s ) exp ( x ) = ( y x ) 1 [ exp ( y x ) I ] exp ( x ) = ( y x ) 1 [ exp ( y ) exp ( x ) ] .

Let f(λ)= n = 0 α n λ n be a function defined by power series with complex coefficients and convergent on the open disk D(0,R)C, R>0 and x,yB with xy=yx. In this paper we establish some upper bounds for the norm of the Čebyšev type difference

f(λ)f(λxy)f(λx)f(λy)
(2.1)

provided that the complex number λ and the vectors x,yB are such that the series in (2.1) are convergent. Applications for some fundamental functions such as the exponential function and the resolvent function are provided as well.

Inequalities for functions of operators in Hilbert spaces may be found in [2326], the recent monographs [2729], and the references therein.

3 The results

We denote by the set of all complex numbers. Let α n be nonzero complex numbers and let

R:= 1 lim sup | α n | 1 n .

Clearly 0R, but we consider only the case 0<R.

Denote by

D(0,R)={ { z C : | z | < R } , if  R < , C , if  R = ,

consider the functions

λf(λ):D(0,R)C,f(λ):= n = 0 α n λ n

and

λ f A (λ):D(0,R)C, f A (λ):= n = 0 | α n | λ n .

Let be a unital Banach algebra and 1 its unity. Denote by

B(0,R)={ { x B : x < R } , if  R < , B , if  R = .

We associate to f the map

x f ˜ (x):B(0,R)B, f ˜ (x):= n = 0 α n x n .

Obviously, f ˜ is correctly defined because the series n = 0 α n x n is absolutely convergent, since n = 0 α n x n n = 0 | α n | x n .

In addition, we assume that s 2 := n = 0 n 2 | α n |<. Let s 0 := n = 0 | α n |< and s 1 := n = 0 n| α n |<.

With the above assumptions we have the following.

Theorem 1 Let λC such that max{|λ|,|λ | 2 }<R< and let x,yB with x,y1, and xy=yx. Then:

  1. (i)

    We have

    f ˜ ( λ 1 ) f ˜ ( λ x y ) f ˜ ( λ x ) f ˜ ( λ y ) 2 ψ min { x 1 , y 1 } f A ( | λ | 2 )
    (3.1)

where

ψ 2 := s 0 s 2 s 1 2 .
(3.2)
  1. (ii)

    We also have

    f ˜ ( λ 1 ) f ˜ ( λ x y ) f ˜ ( λ x ) f ˜ ( λ y ) 2 min { x 1 , y 1 } f A ( | λ | ) × { f A ( | λ | ) [ | λ | f A ( | λ | ) + | λ | 2 f A ′′ ( | λ | ) ] [ | λ | f A ( | λ | ) ] 2 } 1 / 2 .
    (3.3)

Proof For m1 and since xy=yx we have

n = 0 m j = 0 m α n α j λ n λ j ( x n x j ) y n = n = 0 m j = 0 m α n α j λ n λ j x n y n n = 0 m j = 0 m α n α j λ n λ j x j y n = j = 0 m α j λ j n = 0 m α n λ n x n y n j = 0 m α j λ j x j n = 0 m α n λ n y n = j = 0 m α j λ j n = 0 m α n λ n ( x y ) n j = 0 m α j λ j x j n = 0 m α n λ n y n
(3.4)

for any λC.

Taking the norm in (3.4) we have

j = 0 m α j λ j n = 0 m α n λ n ( x y ) n j = 0 m α j λ j x j n = 0 m α n λ n y n n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j ( x n x j ) y n n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j x n x j y n n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j x n x j y n n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j x n x j = 2 0 j < n m | α n | | α j | | λ | n | λ | j x n x j ,
(3.5)

for any λC and m1.

Observe that

L : = 0 j < n m | α n | | α j | | λ | n | λ | j x n x j = 0 j < n m | α n | | α j | | λ | n | λ | j = j n 1 ( x + 1 x ) = 0 j < n m | α n | | α j | | λ | n | λ | j = j n 1 x ( x 1 ) x 1 0 j < n m | α n | | α j | | λ | n | λ | j = j n 1 x
(3.6)

for any λC and m1.

We have

= j n 1 x (nj) max { j , , n 1 } x (nj) max { 1 , , m 1 } x

and then

Lx1 max { 1 , , m 1 } x 0 j < n m | α n || α j ||λ | n |λ | j (nj).
(3.7)

From the first inequality in (3.7) and since x<1 we have

j = 0 m α j λ j n = 0 m α n λ n ( x y ) n j = 0 m α j λ j x j n = 0 m α n λ n y n 2 x 1 0 j < n m | α n | | α j | | λ | n | λ | j ( n j ) = x 1 n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j | n j | .
(3.8)
  1. (i)

    Using the Cauchy-Bunyakovsky-Schwarz inequality for double sums,

    n = 0 m j = 0 m p n , j a n , j b n , j ( n = 0 m j = 0 m p n , j a n , j 2 ) 1 / 2 ( n = 0 m j = 0 m p n , j b n , j 2 ) 1 / 2 ,

where p n , j , a n , j , b n , j 0 for n,j{0,,m}, we have

n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j | n j | ( n = 0 m j = 0 m | α n | | α j | | λ | 2 n | λ | 2 j ) 1 / 2 ( n = 0 m j = 0 m | α n | | α j | | n j | 2 ) 1 / 2 = 2 ( n = 0 m | α n | | λ | 2 n ) [ n = 0 m | α n | n = 0 m n 2 | α n | ( n = 0 m n | α n | ) 2 ] 1 / 2
(3.9)

for any λC and m1.

From (3.8) and (3.9) we get the inequality

j = 0 m α j λ j n = 0 m α n λ n ( x y ) n j = 0 m α j λ j x j n = 0 m α n λ n y n 2 x 1 ( n = 0 m | α n | | λ | 2 n ) × [ n = 0 m | α n | n = 0 m n 2 | α n | ( n = 0 m n | α n | ) 2 ] 1 / 2 .
(3.10)

Since the series

j = 0 α j λ j , n = 0 α n λ n ( x y ) n , j = 0 α j λ j x j , n = 0 α n λ n y n

are convergent in , n = 0 | α n ||λ | 2 n is convergent and the limit

lim m [ n = 0 m | α n | n = 0 m n 2 | α n | ( n = 0 m n | α n | ) 2 ] 1 / 2

exists, then by letting m in (3.10) we deduce the desired result in (3.1) for x. Due to the commutativity of x with y, a similar result can be stated for y, and taking the minimum, we deduce the desired result.

  1. (ii)

    Using the Cauchy-Bunyakovsky-Schwarz inequality for double sums,

    n = 0 m j = 0 m p n , j a n , j ( n = 0 m j = 0 m p n , j ) 1 / 2 ( n = 0 m j = 0 m p n , j a n , j 2 ) 1 / 2

where p n , j , a n , j 0 for n,j{0,,m}, we also have

n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j | n j | ( n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j ) 1 / 2 ( n = 0 m j = 0 m | α n | | α j | | λ | n | λ | j | n j | 2 ) 1 / 2 = 2 ( n = 0 m | α n | | λ | n ) × [ n = 0 m | α n | | λ | n n = 0 m n 2 | α n | | λ | n ( n = 0 m n | α n | | λ | n ) 2 ] 1 / 2
(3.11)

for any λC and m1.

From (3.8) and (3.11) we have

j = 0 m α j λ j n = 0 m α n λ n ( x y ) n j = 0 m α j λ j x j n = 0 m α n λ n y n 2 x 1 ( n = 0 m | α n | | λ | n ) × [ n = 0 m | α n | | λ | n n = 0 m n 2 | α n | | λ | n ( n = 0 m n | α n | | λ | n ) 2 ] 1 / 2
(3.12)

for any λC and m1.

If we denote f(u):= n = 0 α n u n , then for |u|<R we have

n = 0 n α n u n =u f (u)

and

n = 0 n 2 α n u n =u ( u g ( u ) ) .

However

u ( u g ( u ) ) =u g (u)+ u 2 g (u)

and then

n = 0 n 2 α n u n =u g (u)+ u 2 g (u).

Therefore

n = 0 n 2 | α n ||λ | n =|λ| f A ( | λ | ) +|λ | 2 f A ′′ ( | λ | )

and

n = 0 m n| α n ||λ | n =|λ| f ( | λ | )

for |λ|<R.

Since all the series whose partial sums are involved in (3.12) are convergent, then by letting m in (3.12) we deduce the desired inequality (3.3) for x. Due to the commutativity of x with y, a similar result can be stated for y, and taking the minimum, we deduce the desired result. □

Remark 1 If R=, Theorem 1 holds true. Moreover, in this case the restrictions x,y1 need no longer be imposed.

Remark 2 We observe that if the power series f(λ)= n = 0 α n λ n has the radius of convergence R>1, then

n = 0 | α n |= f A (1), n = 0 n 2 | α n |= f A (1)+ f A ′′ (1)

and

n = 0 n| α n |= f A (1).

In this case ψ is finite and

ψ = lim m [ n = 0 m | α n | n = 0 m n 2 | α n | ( n = 0 m n | α n | ) 2 ] 1 / 2 = { f A ( 1 ) [ f A ( 1 ) + f A ′′ ( 1 ) ] [ f A ( 1 ) ] 2 } 1 / 2 .

Therefore, if λC with |λ|,|λ | 2 ,|λ|x,|λ|y<R, then from (3.1) we have

f ˜ ( λ 1 ) f ˜ ( λ x y ) f ˜ ( λ x ) f ˜ ( λ y ) 2 { f A ( 1 ) [ f A ( 1 ) + f A ′′ ( 1 ) ] [ f A ( 1 ) ] 2 } 1 / 2 × min { x 1 , y 1 } f A ( | λ | 2 ) .
(3.13)

Corollary 1 Under the assumptions of Theorem  1 we have the inequalities

f ˜ ( λ 1 ) f ˜ ( λ x 2 ) f ˜ 2 ( λ x ) 2 ψx1 f A ( | λ | 2 )
(3.14)

provided λC with |λ|,|λ | 2 ,|λ|x<R, and

f ˜ ( λ 1 ) f ˜ ( λ x 2 ) f ˜ 2 ( λ x ) 2 x 1 f A ( | λ | ) × { f A ( | λ | ) [ | λ | f A ( | λ | ) + | λ | 2 f A ′′ ( | λ | ) ] [ | λ | f A ( | λ | ) ] 2 } 1 / 2
(3.15)

provided λC with |λ|,|λ|x<R.

Theorem 2 Let f(λ)= n = 0 α n λ n be a function defined by power series with complex coefficients and convergent on the open disk D(0,R)C, R>0, and x,yB with xy=yx and x,y<1.

If λC with |λ|,|λ|x,|λ|y<R, then

f ˜ ( λ 1 ) f ˜ ( λ x y ) f ˜ ( λ x ) f ˜ ( λ y ) min { x 1 1 x , y 1 1 y } [ f A 2 ( | λ | ) f A 2 ( | λ | 2 ) ] ,
(3.16)

where

f A 2 (λ):= n = 0 | α n | 2 λ n
(3.17)

has the radius of convergence R 2 .

Proof As pointed out in (3.6), we have

L x 1 0 j < n m | α n | | α j | | λ | n | λ | j = j n 1 x x 1 = 0 m 1 x 0 j < n m | α n | | α j | | λ | n | λ | j
(3.18)

for any λC and m1.

Denote

K m := 0 j < n m | α n || α j ||λ | n |λ | j .

We obviously have

K m = 1 2 ( n , j = 0 m | α n | | α j | | λ | n | λ | j n = 0 m | α n | 2 | λ | 2 n ) = 1 2 [ ( n = 0 m | α n | | λ | n ) 2 n = 0 m | α n | 2 | λ | 2 n ] .

From (3.8) and (3.18) we get the inequality

j = 0 m α j λ j n = 0 m α n λ n ( x y ) n j = 0 m α j λ j x j n = 0 m α n λ n y n x 1 = 0 m 1 x × [ ( n = 0 m | α n | | λ | n ) 2 n = 0 m | α n | 2 | λ | 2 n ] ,
(3.19)

for any λC and m1.

Since all the series whose partial sums are involved in (3.19) are convergent, then by letting m in (3.19) we deduce the desired inequality (3.16) for x. Due to the commutativity of x with y, a similar result can be stated for y, and taking the minimum, we deduce the desired result. □

Remark 3 Since the power series f A 2 (λ):= n = 0 | α n | 2 λ n is not easy to compute, we can provide some bounds for the quantity

D f ( | λ | ) := f A 2 ( | λ | ) f A 2 ( | λ | 2 ) ,

where |λ|<R, as follows.

If |λ|<1 and a := sup n N {| a n |}<, then

K m a 2 0 j < n m | λ | n | λ | j = 1 2 a 2 [ ( n = 0 m | λ | n ) 2 n = 0 m | λ | 2 n ]

and by taking m in this inequality we get

D f ( | λ | ) 1 2 a 2 [ ( 1 1 | λ | ) 2 1 1 | λ | 2 ]
(3.20)

for |λ|<1.

If |λ|<1 and

a 1 := lim m [ ( n = 0 m | α n | ) 2 n = 0 m | α n | 2 ] <

then

K m max n { 0 , , m } | λ | 2 n 0 j < n m | α n | | α j | 1 2 [ ( n = 0 m | α n | ) 2 n = 0 m | α n | 2 ]

and by taking m in this inequality we get

D f ( | λ | ) 1 2 a 1
(3.21)

for |λ|<1.

If the series n = 0 | α n | and n = 0 | α n | 2 are convergent, then

D f ( | λ | ) 1 2 [ ( n = 0 | α n | ) 2 n = 0 | α n | 2 ]
(3.22)

for |λ|<1.

If |λ|<1, p,q>1 with 1 p + 1 q =1, and

a q := lim m [ ( n = 0 m | α n | q ) 2 n = 0 m | α n | 2 q ] <

then by Hölder’s inequality we have

K m ( 0 j < n m | α n | q | α j | q ) 1 / q ( 0 j < n m | λ | p n | λ | p j ) 1 / p { 1 2 [ ( n = 0 m | α n | q ) 2 n = 0 m | α n | 2 q ] } 1 / q × { 1 2 [ ( n = 0 m | λ | p n ) 2 n = 0 m | λ | 2 p n ] } 1 / p

and by taking m in this inequality we get

D f ( | λ | ) 1 2 a q 1 / q [ ( 1 1 | λ | p ) 2 1 1 | λ | 2 p ] 1 / p
(3.23)

for |λ|<1.

If the series n = 0 | α n | q and n = 0 | α n | 2 q are convergent, then

D f ( | λ | ) 1 2 [ ( n = 0 | α n | q ) 2 n = 0 | α n | 2 q ] 1 / p [ ( 1 1 | λ | p ) 2 1 1 | λ | 2 p ] 1 / p
(3.24)

for |λ|<1.

The following result also holds.

Theorem 3 Let f(λ)= n = 0 α n λ n be a function defined by power series with complex coefficients and convergent on the open disk D(0,R)C, R>0, and x,yB with xy=yx and x,y<1.

If p,q>1 with 1 p + 1 q =1 and λC with |λ|,|λ | p ,|λ|x,|λ|y<R, then

f ˜ ( λ 1 ) f ˜ ( λ x y ) f ˜ ( λ x ) f ˜ ( λ y ) 1 2 min { x 1 ( 1 x p ) 1 / p , y 1 ( 1 y p ) 1 / p } × φ 1 / q [ f A 2 ( | λ | p ) f A 2 ( | λ | 2 p ) ] 1 / p ,
(3.25)

where

φ:= lim m n , j = 0 m | α n || α j ||nj|
(3.26)

is assumed to exist and be finite.

Proof Using Hölder’s inequality for p,q>1 with 1 p + 1 q =1 and (3.6), we have

L x 1 0 j < n m | α n | | α j | | λ | n | λ | j ( n j ) 1 / q ( = j n 1 x p ) 1 / p x 1 ( = 0 m 1 x p ) 1 / p 0 j < n m | α n | | α j | | λ | n | λ | j ( n j ) 1 / q
(3.27)

for any λC and m1.

Applying Hölder’s inequality once more we have

0 j < n m | α n | | α j | | λ | n | λ | j ( n j ) 1 / q ( 0 j < n m | α n | | α j | | λ | n ( n j ) ) 1 / q ( 0 j < n m | α n | | α j | | λ | p n | λ | p j ) 1 / p = ( 1 2 n , j = 0 m | α n | | α j | | n j | ) 1 / q × ( 1 2 [ ( n = 0 m | α n | | λ | n p ) 2 n = 0 m | α n | 2 | λ | 2 n p ] ) 1 / p = 1 2 ( n , j = 0 m | α n | | α j | | n j | ) 1 / q × [ ( n = 0 m | α n | | λ | n p ) 2 n = 0 m | α n | 2 | λ | 2 n p ] 1 / p
(3.28)

for any λC and m1.

From (3.8) and (3.28) we get the inequality

j = 0 m α j λ j n = 0 m α n λ n ( x y ) n j = 0 m α j λ j x j n = 0 m α n λ n y n 1 2 x 1 ( = 0 m 1 x p ) 1 / p ( n , j = 0 m | α n | | α j | | n j | ) 1 / q × [ ( n = 0 m | α n | | λ | n p ) 2 n = 0 m | α n | 2 | λ | 2 n p ] 1 / p ,
(3.29)

for any λC and m1.

Since all the series whose partial sums are involved in (3.29) are convergent, then by letting m in (3.29) we deduce the desired inequality (3.25) for x. Due to the commutativity of x with y, a similar result can be stated for y, and taking the minimum, we deduce the desired result. □

Remark 4 Observe that

[ f A 2 ( | λ | p ) f A 2 ( | λ | 2 p ) ] 1 / p = D f 1 / p ( | λ | p )

and then further bounds for the inequality (3.25) may be provided by the use of Remark 3. However the details are not mentioned here.

We can obtain a simpler upper bound for φ as follows.

Using the Cauchy-Bunyakovsky-Schwarz inequality for double sums

n = 0 m j = 0 m p i , j a i , j ( n = 0 m j = 0 m p i , j ) 1 / 2 ( n = 0 m j = 0 m p i , j a i , j 2 ) 1 / 2 ,

where p i , j , a i , j 0 for i,j{0,,m}, we have

n , j = 0 m | α n | | α j | | n j | ( n , j = 0 m | α n | | α j | ) 1 / 2 ( n , j = 0 m | α n | | α j | | n j | 2 ) 1 / 2 = 2 n = 0 m | α n | [ n = 0 m | α n | n = 0 m n 2 | α n | ( n = 0 m n | α n | ) 2 ] 1 / 2
(3.30)

for m1.

If the series n = 0 | α n | is finite and ψ, defined by (3.2), is finite, then from (3.30) we have

φ 2 n = 0 | α n |ψ.
(3.31)

We observe that, if the power series f(λ)= n = 0 α n λ n has the radius of convergence R>1, then ψ is finite and

ψ= { f A ( 1 ) [ f A ( 1 ) + f A ′′ ( 1 ) ] [ f A ( 1 ) ] 2 } 1 / 2 .

We have from (3.31) the inequality

φ 2 f A (1) { f A ( 1 ) [ f A ( 1 ) + f A ′′ ( 1 ) ] [ f A ( 1 ) ] 2 } 1 / 2 .
(3.32)

4 Some examples

As some natural examples that are useful for applications, we can point out that, if

f ( λ ) = n = 1 ( 1 ) n n λ n = ln 1 1 + λ , λ D ( 0 , 1 ) ; g ( λ ) = n = 0 ( 1 ) n ( 2 n ) ! λ 2 n = cos λ , λ C ; h ( λ ) = n = 0 ( 1 ) n ( 2 n + 1 ) ! λ 2 n + 1 = sin λ , λ C ; l ( λ ) = n = 0 ( 1 ) n λ n = 1 1 + λ , λ D ( 0 , 1 ) ,
(4.1)

then the corresponding functions constructed by the use of the absolute values of the coefficients are

f A ( λ ) = n = 1 1 n λ n = ln 1 1 λ , λ D ( 0 , 1 ) ; g A ( λ ) = n = 0 1 ( 2 n ) ! λ 2 n = cosh λ , λ C ; h A ( λ ) = n = 0 1 ( 2 n + 1 ) ! λ 2 n + 1 = sinh λ , λ C ; l A ( λ ) = n = 0 λ n = 1 1 λ , λ D ( 0 , 1 ) .
(4.2)

Other important examples of functions as power series representations with nonnegative coefficients are

exp ( λ ) = n = 0 1 n ! λ n , λ C , 1 2 ln ( 1 + λ 1 λ ) = n = 1 1 2 n 1 λ 2 n 1 , λ D ( 0 , 1 ) ; sin 1 ( λ ) = n = 0 Γ ( n + 1 2 ) π ( 2 n + 1 ) n ! λ 2 n + 1 , λ D ( 0 , 1 ) ; tanh 1 ( λ ) = n = 1 1 2 n 1 λ 2 n 1 , λ D ( 0 , 1 ) ; F 1 2 ( α , β , γ , λ ) = n = 0 Γ ( n + α ) Γ ( n + β ) Γ ( γ ) n ! Γ ( α ) Γ ( β ) Γ ( n + γ ) λ n , α , β , γ > 0 , λ D ( 0 , 1 ) ;
(4.3)

where Γ is the Gamma function.

If we apply the inequality (3.13) to the exponential function, then we have

exp [ λ ( 1 + x y ) ] exp [ λ ( x + y ) ] 2 emin { x 1 , y 1 } exp ( | λ | 2 )
(4.4)

for any x,yB with xy=yx, x,y<1, and λC.

If we take y=x in (4.4), then we get

exp [ λ ( 1 x 2 ) ] 1 2 emin { x 1 , x + 1 } exp ( | λ | 2 )
(4.5)

for any xB with x<1 and λC.

If we apply the inequality (3.3) for the exponential functions we also have

exp [ λ ( 1 + x y ) ] exp [ λ ( x + y ) ] 2 min { x 1 , y 1 } | λ | 1 / 2 exp ( 2 | λ | ) ,
(4.6)

for any x,yB with xy=yx, x,y<1, and λC.

If we take y=x in (4.6), then we get

exp [ λ ( 1 x 2 ) ] 1 2 min { x 1 , x + 1 } |λ | 1 / 2 exp ( 2 | λ | ) .
(4.7)

Now, consider the function f(λ):= n = 0 λ n = 1 1 λ , λD(0,1). If we apply the inequality (3.3) for this function, then we get the result

( 1 λ ) 1 ( 1 λ x y ) 1 ( 1 λ x ) 1 ( 1 λ y ) 1 2 min { x 1 , y 1 } | λ | 1 / 2 ( 1 | λ | ) 3
(4.8)

for any x,yB with xy=yx, x,y<1, and λC with |λ|<1.

We have in particular the inequalities

( 1 λ ) 1 ( 1 λ x 2 ) 1 ( 1 λ x ) 2 2 x1|λ | 1 / 2 ( 1 | λ | ) 3
(4.9)

and

( 1 λ ) 1 ( 1 + λ x 2 ) 1 ( 1 λ 2 x 2 ) 1 2 min { x 1 , x + 1 } | λ | 1 / 2 ( 1 | λ | ) 3
(4.10)

for any xB with x<1 and λC with |λ|<1.

Now, if we take λ= 1 γ with |γ|>1 then we get from (4.8) the inequality

γ 2 ( γ 1 ) 1 ( γ x y ) 1 γ 2 ( γ x ) 1 ( γ y ) 1 2 min { x 1 , y 1 } | γ | 1 / 2 ( | γ | 1 ) 3 | γ | 3 ,

which is equivalent with

( γ 1 ) 1 ( γ x y ) 1 ( γ x ) 1 ( γ y ) 1 2 min { x 1 , y 1 } | γ | 1 / 2 ( | γ | 1 ) 3

for any x,yB with xy=yx, x,y<1, and γC with |γ|>1.

If we use the resolvent function notation, then we have the following inequality:

( γ 1 ) 1 R x y ( γ ) R x ( γ ) R y ( γ ) 2 min { x 1 , y 1 } | γ | 1 / 2 ( | γ | 1 ) 3
(4.11)

for any x,yB with xy=yx, x,y<1, and γC with |γ|>1.

In particular, we have

( γ 1 ) 1 R x 2 ( γ ) R x 2 ( γ ) 2 x1|γ | 1 / 2 ( | γ | 1 ) 3
(4.12)

for any xB with x<1 and γC with |γ|>1.

Remark 5 Similar inequalities may be stated for the other power series mentioned at the beginning of this paragraph. However, the details are not presented here.

References

  1. Grüss G:Über das Maximum des absoluten Betrages von 1 b a a b f(x)g(x)dx 1 ( b a ) 2 a b f(x)dx a b g(x)dx. Math. Z. 1935, 39: 215–226. 10.1007/BF01201355

    MathSciNet  Article  MATH  Google Scholar 

  2. Chebyshev PL: Sur les expressions approximatives des intégrals définies par les autres prises entre les mêmes limites. Proc. Math. Soc. Charkov 1882, 2: 93–98.

    Google Scholar 

  3. Ostrowski AM: On an integral inequality. Aequ. Math. 1970, 4: 358–373. 10.1007/BF01844168

    MathSciNet  Article  MATH  Google Scholar 

  4. Lupaş A: The best constant in an integral inequality. Mathematica 1973,15(38)(2):219–222.

    MathSciNet  MATH  Google Scholar 

  5. Cerone P, Dragomir SS: New bounds for the Čebyšev functional. Appl. Math. Lett. 2005, 18: 603–611. 10.1016/j.aml.2003.09.013

    MathSciNet  Article  MATH  Google Scholar 

  6. Cheng X-L, Sun J: Note on the perturbed trapezoid inequality. J. Inequal. Pure Appl. Math. 2002.,3(2): Article ID 29. http://www.emis.de/journals/JIPAM/article181.html?sid=181

  7. Cerone P, Dragomir SS: A refinement of the Grüss inequality and applications. Tamkang J. Math. 2007,38(1):37–49. Preprint, RGMIA Res. Rep. Coll. 5(2), Art. 14 (2002). http://rgmia.org/papers/v5n2/RGIApp.pdf

    MathSciNet  MATH  Google Scholar 

  8. Alomari MW: Some Grüss type inequalities for Riemann-Stieltjes integral and applications. Acta Math. Univ. Comen. 2012,81(2):211–220.

    MathSciNet  MATH  Google Scholar 

  9. Dragomir SS: Some inequalities for Riemann-Stieltjes integral and applications. In Optimisation and Related Topics. Edited by: Rubinov A. Kluwer Academic, Dordrecht; 2000:197–235.

    Google Scholar 

  10. Dragomir SS: Sharp bounds of Čebyšev functional for Stieltjes integrals and applications. Bull. Aust. Math. Soc. 2003,67(2):257–266. 10.1017/S0004972700033724

    Article  MATH  Google Scholar 

  11. Dragomir SS: New estimates of the Čebyšev functional for Stieltjes integrals and applications. J. Korean Math. Soc. 2004,41(2):249–264.

    MathSciNet  Article  MATH  Google Scholar 

  12. Dragomir SS: Čebyšev’s type inequalities for functions of selfadjoint operators in Hilbert spaces. Linear Multilinear Algebra 2010,58(7–8):805–814.

    MathSciNet  Article  MATH  Google Scholar 

  13. Dragomir SS: Inequalities for the Čebyšev functional of two functions of selfadjoint operators in Hilbert spaces. Aust. J. Math. Anal. Appl. 2009,6(1):1–58.

    MathSciNet  MATH  Google Scholar 

  14. Dragomir SS: Some inequalities for power series of selfadjoint operators in Hilbert spaces via reverses of the Schwarz inequality. Integral Transforms Spec. Funct. 2009,20(9–10):757–767.

    MathSciNet  Article  MATH  Google Scholar 

  15. Gavrea B: Improvement of some inequalities of Chebysev-Grüss type. Comput. Math. Appl. 2012,64(6):2003–2010. 10.1016/j.camwa.2012.03.101

    MathSciNet  Article  MATH  Google Scholar 

  16. Li X, Mohapatra RN, Rodriguez RS: Grüss-type inequalities. J. Math. Anal. Appl. 2002,267(2):434–443. 10.1006/jmaa.2001.7565

    MathSciNet  Article  MATH  Google Scholar 

  17. Mercer AM: An improvement of the Grüss inequality. J. Inequal. Pure Appl. Math. 2005.,6(4): Article ID 93

  18. Otachel Z: Grüss type inequalities in normed spaces. J. Math. Anal. Appl. 2012,393(1):14–24. 10.1016/j.jmaa.2012.04.004

    MathSciNet  Article  MATH  Google Scholar 

  19. Pachpatte BG: On Grüss like integral inequalities via Pompeiu’s mean value theorem. J. Inequal. Pure Appl. Math. 2005.,6(3): Article ID 82

  20. Dragomir SS, Pečarić J, Sándor J: The Chebyshev inequality in pre-Hilbertian spaces. II. In Proceedings of the Third Symposium of Mathematics and Its Applications, Timişoara, 1989. Rom. Acad, Timişoara; 1990:75–78. MR1266442 (94m:46033)

    Google Scholar 

  21. Dragomir SS, Sándor J: The Chebyshev inequality in pre-Hilbertian spaces. I. In Proceedings of the Second Symposium of Mathematics and Its Applications, Timişoara, 1987. Res. Centre, Acad. SR, Romania, Timişoara; 1988:61–64. MR1006000 (90k:46048)

    Google Scholar 

  22. Liu Z: Refinement of an inequality of Grüss type for Riemann-Stieltjes integral. Soochow J. Math. 2004,30(4):483–489.

    MathSciNet  MATH  Google Scholar 

  23. Cohen JE: Chebyshev and Grüss inequalities for real rectangular matrices. Linear Algebra Appl. 2014, 447: 133–138.

    MathSciNet  Article  MATH  Google Scholar 

  24. Jocić DR, Krtinić D, Moslehian MS: Landau and Grüss type inequalities for inner product type integral transformers in norm ideals. Math. Inequal. Appl. 2013,16(1):109–125.

    MathSciNet  MATH  Google Scholar 

  25. Matharu JS, Aujla JS: Hadamard product versions of the Chebyshev and Kantorovich inequalities. J. Inequal. Pure Appl. Math. 2009.,10(2): Article ID 51

  26. Moslehian MS, Bakherad M: Chebyshev type inequalities for Hilbert space operators. J. Math. Anal. Appl. 2014. 10.1016/j.jmaa.2014.05.078

    Google Scholar 

  27. Dragomir SS Springer Briefs in Mathematics. In Operator Inequalities of the Jensen, Čebyšev and Grüss Type. Springer, New York; 2012.

    Chapter  Google Scholar 

  28. Dragomir SS Springer Briefs in Mathematics. In Operator Inequalities of Ostrowski and Trapezoidal Type. Springer, New York; 2012.

    Chapter  Google Scholar 

  29. Furuta T, Mićić Hot J, Pečarić J, Seo Y Inequalities for Bounded Selfadjoint Operators on a Hilbert Space. In Mond-Pečarić Method in Operator Inequalities. Element, Zagreb; 2005.

    Google Scholar 

Download references

Acknowledgements

The authors would like to thank the anonymous referees for their valuable comments that have been implemented in the final version of the paper.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Silvestru S Dragomir.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

All authors contributed equally and significantly in writing this article. All authors read and approved the final manuscript.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0), which permits use, duplication, adaptation, distribution, and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Dragomir, S.S., Boldea, M.V., Buşe, C. et al. Norm inequalities of Čebyšev type for power series in Banach algebras. J Inequal Appl 2014, 294 (2014). https://doi.org/10.1186/1029-242X-2014-294

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/1029-242X-2014-294

Keywords

  • Banach algebras
  • power series
  • exponential function
  • resolvent function
  • norm inequalities