In this paper converses of a generalized Jensen’s inequality for a continuous field of self-adjoint operators, a unital field of positive linear mappings and real-valued continuous convex functions are studied. New refined converses are presented by using the Mond-Pečarić method improvement. Obtained results are applied to refine selected inequalities with power functions.
The online version of this article (doi:10.1186/1029-242X-2013-353) contains supplementary material, which is available to authorized users.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors contributed equally and significantly in writing this paper. All authors read and approved the final manuscript.
1 Introduction
Let T be a locally compact Hausdorff space and let be a -algebra of operators on some Hilbert space H. We say that a field of operators in is continuous if the function is norm continuous on T. If in addition μ is a Radon measure on T and the function is integrable, then we can form the Bochner integral , which is the unique element in such that
for every linear functional φ in the norm dual .
Anzeige
Assume further that there is a field of positive linear mappings from to another -algebra ℬ of operators on a Hilbert space K. We recall that a linear mapping is said to be positive if for all . We say that such a field is continuous if the function is continuous for every . Let the -algebras include the identity operators and let the function be integrable with for some positive scalar k. If , we say that a field is unital.
Let be the -algebra of all bounded linear operators on a Hilbert space H. We define bounds of a self-adjoint operator by
(1)
for . If denotes the spectrum of x, then .
For an operator , we define the operator . Obviously, if x is self-adjoint, then .
Anzeige
Jensen’s inequality is one of the most important inequalities. It has many applications in mathematics and statistics and some other well-known inequalities are its special cases.
Let f be an operator convex function defined on an interval I. Davis [1] proved the so-called Jensen operator inequality
(2)
where is a unital completely positive linear mapping from a -algebra to linear operators on a Hilbert space K, and x is a self-adjoint element in with spectrum in I. Subsequently, Choi [2] noted that it is enough to assume that ϕ is unital and positive.
Mond, Pečarić, Hansen, Pedersen et al. in [3‐6] studied another generalization of (2) for operator convex functions. Moreover, Hansen et al. [7] presented a general formulation of Jensen’s operator inequality for a bounded continuous field of self-adjoint operators and a unital field of positive linear mappings:
(3)
where f is an operator convex function.
There is an extensive literature devoted to Jensen’s inequality concerning different refinements and extensive results, e.g., see [8‐20]. Mićić et al. [21] proved that the discrete version of (3) stands without operator convexity of f under a condition on the spectra of operators. Recently, Mićić et al. [22] presented a discrete version of refined Jensen’s inequality for real-valued continuous convex functions. A continuous version is given below.
Theorem 1Letbe a bounded continuous field of self-adjoint elements in a unital -algebradefined on a locally compact Hausdorff spaceTequipped with a bounded Radon measureμ. Letand , , be the bounds of , . Letbe a unital field of positive linear mappingsfromto another unital -algebra ℬ. Let
whereand , , are the bounds of the operatorand
Ifis a continuous convex (resp. concave) function provided that the intervalIcontains all , , then
(resp.
(4)
holds, where
and , , , are arbitrary numbers.
The proof is similar to [[22], Theorem 3] and we omit it.
On the other hand, Mond, Pečarić, Furuta et al. in [6, 23‐27] investigated converses of Jensen’s inequality. For presenting these results, we introduce some abbreviations. Let , . Then a linear function through and has the form , where
(5)
Using the Mond-Pečarić method, in [27] the following generalized converse of Jensen’s operator inequality (2) is presented
(6)
for a convex function f defined on an interval , , where g is a real-valued continuous function on , is a real-valued function defined on , operator monotone in u, , , is a unital positive linear mapping and A is a self-adjoint operator with spectrum contained in .
A continuous version of (6) and in the case of for some positive scalar k, is presented in [28]. Recently, Mićić et al. [29] obtained better bound than the one given in (6) as follows.
Letbe a bounded continuous field of self-adjoint elements in a unital -algebrawith the spectra in , , defined on a locally compact Hausdorff spaceTequipped with a bounded Radon measureμ, and letbe a unital field of positive linear mapsfromto another unital -algebra ℬ. Letand , , be the bounds of the self-adjoint operatorand , , , where , andFis bounded.
Iffis convex andFis an operator monotone in the first variable, then
(7)
where constantsandare
Iffis concave, then reverse inequalities are valid in (7) with inf instead of sup in boundsandC.
In this paper, we present refined converses of Jensen’s operator inequality. Applying these results, we further refine selected inequalities with power functions.
2 Main results
In the following we assume that is a bounded continuous field of self-adjoint elements in a unital -algebra with the spectra in , , defined on a locally compact Hausdorff space T equipped with a bounded Radon measure μ and that is a unital field of positive linear mappings between -algebras.
For convenience, we introduce abbreviations and as follows:
(8)
where m, M, , are some scalars such that the spectra of , , are in ;
(9)
where is a continuous function.
Obviously, implies for and . It follows . Also, if f is convex (resp. concave), then (resp. ).
To prove our main result related to converse Jensen’s inequality, we need the following lemma.
Lemma 3Letfbe a convex function on an intervalI, andsuch that . Then
(10)
Proof These results follow from [[30], Theorem 1, p.717] for . For the reader’s convenience, we give an elementary proof of (10).
Let , , be positive real numbers such that . Using Jensen’s inequality and its reverse, we get
(11)
Suppose that , . Replacing and by and , respectively, and putting , and in (11), we get
which gives the right-hand side of (10). Similarly, replacing and by and , respectively, and putting , and in (11), we obtain the left-hand side of (10).
If , or , , then inequality (10) holds, since f is convex. If , then we have an equality in (10). □
The main result of an improvement of the Mond-Pečarić method follows.
Lemma 4Let , , mandMbe as above. Then
(12)
for every continuous convex function , whereandare defined by (8) and (9), respectively.
Iffis concave, then the reverse inequality is valid in (12).
Proof We prove only the convex case. By using (10) we get
(13)
for every such that . Let functions be defined by
Then, for any , we can write
By using (13) we get
(14)
where
since
Now since , by utilizing the functional calculus to (14), we obtain
where
Applying a positive linear mapping , integrating and using , we get the first inequality in (12) since
By using that , the second inequality in (12) holds. □
We can use Lemma 4 to obtain refinements of some other inequalities mentioned in the introduction. First, we present a refinement of Theorem 2.
Theorem 5Letand , , be the bounds of the operatorand letbe the lower bound of the operator . Let , , , where , andFis bounded.
Iffis convex andFis operator monotone in the first variable, then
(15)
Iffis concave, then the reverse inequality is valid in (15) with inf instead of sup.
Proof We prove only the convex case. Then implies . By using (12) it follows that
Using operator monotonicity of in the first variable, we obtain (15). □
3 Difference-type converse inequalities
By using Jensen’s operator inequality, we obtain that
(16)
holds for every operator convex function f on , every function g and real number α such that on . Now, applying Theorem 5 to the function , , we obtain the following converse of (16). It is also a refinement of [[29], Theorem 3.1].
Theorem 6Letand , , be the bounds of the operatorand , be continuous functions.
Iffis convex and , then
(17)
Iffis concave, then the reverse inequality is valid in (17) with min instead of max.
Remark 1 (1) Obviously,
for every convex function f, every , and , where is the lower bound of .
(2)
According to [[29], Corollary 3.2], we can determine the constant in the RHS of (17).
(i)
Let f be convex. We can determine the value in
as follows:
if , g is convex or , g is concave, then
(18)
if , g is concave or , g is convex, then
(19)
(ii)
Let f be concave. We can determine the value in
as follows:
if , g is convex or , g is concave, then is equal to the right-hand side in (19) with reverse inequality signs;
if , g is concave or , g is convex, then is equal to the right-hand side in (18) with min instead of max.
Theorem 6 and Remark 1(2) applied to functions and give the following corollary, which is a refinement of [[29], Corollary 3.3].
Corollary 7Letbe a field of strictly positive operators, letand , , be the bounds of the operator . Letbe defined by (8).
(i)
Let . Then
where the constantis determined as follows:
if, or, , then
(20)
if, or, , then
(21)
whereand .
(ii)
Let . Then
where the constantis determined as follows:
if, or, , thenis equal to the right-hand side in (21);
if, or, , thenis equal to the right-hand side in (20) with min instead of max.
Using Theorem 6 and Remark 1 for and and utilizing elementary calculations, we obtain the following converse of Jensen’s inequality.
Theorem 8Letand , , be the bounds of the operatorand letbe a continuous function.
Iffis convex, then
(22)
whereandare defined by (8) and (9), respectively, and
(23)
Furthermore, iffis strictly convex differentiable, then the boundsatisfies the following condition:
whereis the lower bound of the operator . We can determine the valuein (23) as follows:
(24)
where
(25)
In the dual case, whenfis concave, the reverse inequality is valid in (22) with min instead of max in (23). Furthermore, iffis strictly concave differentiable, then the boundsatisfies the following condition:
We can determine the valuein (24) with , which equals the right-hand side in (25) with reverse inequality signs.
Example 1 We give examples for the matrix cases and . We put , which is convex, but not operator convex. Also, we define mappings by , and measures by .
(I)
First, we observe an example without the spectra condition (see Figure 1(a)). Then we obtain a refined inequality as in (22), but do not have refined Jensen’s inequality.
and , , , , , (rounded to three decimal places). We have
and
since , , .
×
(II)
Next, we observe an example with the spectra condition (see Figure 1(b)). Then we obtain a series of inequalities involving refined Jensen’s inequality and its converses.
and , , , , , , , and we put , (rounded to three decimal places). We have
since , , , and .
Applying Theorem 8 to , we obtain the following refinement of [[29], Corollary 3.6].
Corollary 9Letbe a field of strictly positive operators, letand , , be the bounds of the operator . Letbe defined by (8). Then
for , and
for , where
(26)
andequals the right-hand side in (26) with reverse inequality signs. is the known Kantorovich-type constant for difference (see, i.e., [[6], §2.7]):
4 Ratio-type converse inequalities
In [[29], Theorem 4.1] the following ratio-type converse of (16) is given:
(27)
where f is convex and . Applying Theorem 5 and Theorem 6, we obtain the following two refinements of (27).
Theorem 10Letand , , be the bounds of the operatorand let , be continuous functions.
Iffis convex and , then
(28)
and
(29)
whereandare defined by (8) and (9), respectively, andis the lower bound of the operator . Iffis concave, then reverse inequalities are valid in (28) and (29) with min instead of max.
Proof We prove only the convex case. Let . Then there is such that and for all . It follows that and for all . So,
By using (17), we obtain (28). Inequality (29) follows directly from Theorem 5 by putting . □
Remark 2 (1) Inequality (28) is a refinement of (27) since . Also, (29) is a refinement of (27) since and implies
(2)
Let the assumptions of Theorem 10 hold. Generally, there is no relation between the right-hand sides of inequalities (28) and (29) under the operator order (see Example 2). But, for example, if , where is the point where it achieves , then the following order holds:
Example 2 Let , and , .
and , , , , , (rounded to three decimal places). We have
(30)
since , , . Further,
(31)
since . We remark that there is no relation between matrices in the right-hand sides of equalities (30) and (31).
Remark 3 Similar to [[29], Corollary 4.2], we can determine the constant in the RHS of (29).
(i)
Let f be convex. We can determine the value C in
as follows:
if g is convex, then
(32)
if g is concave, then
(33)
Also, we can determine the constant D in
in the same way as the above constant C but without .
(ii)
Let f be concave. We can determine the value c in
as follows:
if g is convex, then c is equal to the right-hand side in (33) with min instead of max;
if g is concave, then c is equal to the right-hand side in (32) with reverse inequality signs.
Also, we can determine the constant d in
in the same way as the above constant c but without .
Theorem 10 and Remark 3 applied to functions and give the following corollary, which is a refinement of [[29], Corollary 4.4].
Corollary 11Letbe a field of strictly positive operators, letand , , be the bounds of the operator . Letbe defined by (8), be the lower bound of the operatorand .
(i)
Let . Then
where the constantis determined as follows:
if, then
(34)
if, then
(35)
Also,
holds, whereis determined in the same way as the above constantbut without .
(ii)
Let . Then
where the constantis determined as follows:
if, thenis equal to the right-hand side in (35) with min instead of max;
if, thenis equal to the right-hand side in (34).
Also,
holds, where , andis determined in the same way as the above constantbut without .
Using Theorem 10 and Remark 3 for and utilizing elementary calculations, we obtain the following converse of Jensen’s operator inequality.
Theorem 12Letand , , be the bounds of the operator .
Ifis a continuous convex function and strictly positive on , then
(36)
and
(37)
whereandare defined by (8) and (9), respectively, andis the lower bound of the operator .
In the dual case, iffis concave, then the reverse inequalities are valid in (36) and (37) with min instead of max.
Furthermore, iffis convex differentiable on , we can determine the constant
in (36) as follows:
(38)
Also, iffis strictly convex twice differentiable on , then we can determine the constant
in (37) as follows:
(39)
whereis defined as the unique solution of the equationprovided . Otherwise, is defined asorprovidedor , respectively.
In the dual case, iffis concave differentiable, then the valueis equal to the right-hand side in (38) with reverse inequality signs. Also, iffis strictly concave twice differentiable, then we can determine the valuein (39) with , which equals the right-hand side in (39) with reverse inequality signs.
Remark 4 If f is convex and strictly negative on , then (36) and (37) are valid with min instead of max. If f is concave and strictly negative, then reverse inequalities are valid in (36) and (37).
Applying Theorem 12 to , we obtain the following refinement of [[29], Corollary 4.8].
Corollary 13Letbe a field of strictly positive operators, letand , , be the bounds of the operator . Letbe defined by (8), be the lower bound of the operatorand .
If , then
(40)
and
(41)
where
(42)
is a generalization of the known Kantorovich constant (defined in [[6], §2.7]) as follows:
(43)
forand .
If , then
and
whereequals the right-hand side in (42) with reverse inequality signs.
Proof The second inequalities in (40) and (41) follow directly from (37) and (36) by using (39) and (38), respectively. The last inequality in (40) follows from
The third inequality in (41) follows from
since for and . □
Appendix 1: A new generalization of the Kantorovich constant
Definition 1 Let . Further generalization of Kantorovich constant (given in [[6], Definition 2.2]) is defined by
for any real number and any c, . The constant is sometimes denoted by briefly. Some of those constants are depicted in Figure 2.
×
By inserting in , we obtain the Kantorovich constant . The constant defined by (43) coincides with by putting .
Lemma 14Let . The generalized Kantorovich constanthas the following properties:
(i)
for all ,
(ii)
for allandfor all ,
(iii)
is decreasing ofcforand increasing ofcfor ,
(iv)
for allandfor all ,
(v)
for all .
Proof (i) We use an easy calculation:
(ii)
Let . The logarithms calculation and l’Hospital’s theorem give as , as and as . Now using (i) we obtain (ii).
(iii)
Let and .
Since the function is convex (resp. concave) on if (resp. ), then (resp. ) for every . Then if and if , which gives that is decreasing of c if and increasing of c if .
(iv)
Let and . If then
implies
which gives . Similarly, if and if . Next, using (iii) and [[6], Theorem 2.54(iv)], for .
(v)
Let . Using (iii) and [[6], Theorem 2.54(vi)], . □
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 2.0 International License (https://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
All authors contributed equally and significantly in writing this paper. All authors read and approved the final manuscript.