Abstract
In this paper, by the definition of Berezin number, we present some inequalities involving the operator geometric mean. For instance, it is shown that if \(X, Y, Z\in {\mathcal{L}}(\mathcal{H})\) such that X and Y are positive operators, then
in which \(X\mathbin{\sharp} Y=X^{\frac{1}{2}} ( X^{-\frac{1}{2}}YX^{- \frac{1}{2}} ) ^{\frac{1}{2}}X^{\frac{1}{2}}\), \(p\geq q>1\) such that \(r\geq \frac{2}{q}\) and \(\frac{1}{p}+\frac{1}{q}=1\).
Similar content being viewed by others
1 Introduction and preliminaries
We denote the \(C^{*}\)-algebra of all bounded linear operators on a separable complex Hilbert space \({\mathcal{H}}\) with \(\mathcal{L} ( \mathcal{H} ) \). An operator \(X\in \mathcal{L} ( \mathcal{H} ) \) is called positive if \(\langle Xx,x\rangle \geq 0\) for every \(x\in {\mathcal{H} }\), and in this case we write \(X\geq 0\). The numerical range and numerical radius of \(X\in \mathcal{L} ( \mathcal{H} ) \) are respectively defined by \(W ( X ) := \{ \langle Xf,f \rangle :f \in \mathcal{H}, \Vert f \Vert =1 \}\) and \(w ( X ) :=\sup \{ \vert f \vert :f \in W ( X ) \} \). We denote by \(\mathcal{F} ( \varOmega ) \) the set of all complex-valued functions on a nonempty set Ω. Let \(\mathcal{H}=\mathcal{H} ( \varOmega ) \subset \mathcal{F} ( \varOmega ) \) be a Hilbert space. The Riesz representation theorem makes certain that a functional Hilbert space has a reproducing kernel, which is a function \(k_{\lambda }:\varOmega \times \varOmega \rightarrow \mathcal{H}\), that is called the reproducing kernel enjoying the reproducing property \(k_{\lambda }:=k ( \cdot, \lambda ) \in \mathcal{H}\) (\(\lambda \in \varOmega \)) such that \(f(\lambda )= \langle f,k_{\lambda } \rangle _{\mathcal{H}}\), in which \(\lambda \in \varOmega \) and \(f\in \mathcal{H}\) (see [18]). For \(\{ \xi _{n} ( z ) \} _{n \geq 0}\), an orthonormal basis of the space \(\mathcal{H} ( \varOmega ) \), the reproducing kernel can be presented as follows:
(see [2, 18] and the references therein). Throughout the paper, \(\mathcal{H}=\mathcal{H}(\varOmega )\) for some nonempty set Ω. If \(X\in \mathcal{L}(\mathcal{H})\), then the Berezin symbol of X is the function X̃ with
where \(\widehat{k}_{\lambda }=\frac{k_{\lambda }}{ \Vert k_{ \lambda } \Vert }\) is the normalized reproducing kernel of \(\mathcal{H}\) (see [7]). Karaev in [13–15] defined the Berezin set and the Berezin number for operator X as follows:
respectively. Moreover, the Berezin number of two operators X, Y satisfies the following properties:
- (i)
\(\operatorname{ber} ( \nu X ) =|\nu | \operatorname{ber} ( X ) \) for all \(\nu \in \mathcal{C}\);
- (ii)
\(\operatorname{ber} ( X+Y ) \leq \operatorname{ber} ( X ) +\operatorname{ber} ( X ) \).
Also, we know that
for all \(X\in \mathcal{L} ( \mathcal{H} ) \). In some recent papers, several Berezin number inequalities have been investigated by authors [3–6, 9, 10, 12, 21, 22].
Assume that \(X_{1},\ldots,X_{n}\in \mathcal{L} ( \mathcal{H} ) \) and \(p\geq 1\). In [3], the generalized Euclidean Berezin number of \(X_{1},\ldots ,X_{n}\) is defined as follows:
If \(p,q>1\) with \(\frac{1}{p}+\frac{1}{q}=1\), then the Young inequality is the inequality
where x and y are positive real numbers (see [11]). A refinement of (1) was obtained by Kittaneh and Manasrah [17]
where \(r_{0}=\min \{ \frac{1}{p},\frac{1}{q} \} \) or equivalently
in which \(\nu \in [0,1]\) and \(r_{0}=\min \{\nu ,1-\nu \}\).
For positive operators \(X,Y\in \mathcal{L} ( \mathcal{H} ) \), the operator geometric mean is the positive operator \(X\mathbin{\sharp} Y=X ^{\frac{1}{2}} ( X^{-\frac{1}{2}}YX^{-\frac{1}{2}} ) ^{ \frac{1}{2}}X^{\frac{1}{2}} \), where it has the property \(X\mathbin{\sharp} Y=Y \mathbin{\sharp} X\). A matrix mean inequality was established by Bhatia and Kittaneh in [8], and later this inequality was generalized in [18]. A matrix Young inequality was obtained by Ando in [1]. The matrix mean inequality and the matrix Young inequality were considered with the numerical radius norm by Salemi and Sheikhhosseini in [19, 20].
In this paper, we get some upper bounds for the Berezin number of the \(( X\mathbin{\sharp} Y ) Z\) on reproducing kernel Hilbert spaces (RKHS), where \(Z\in \mathcal{L} ( \mathcal{H} ) \) is arbitrary, and give some Berezin number inequalities. We also present some inequalities for the generalized Euclidean Berezin number.
2 Main results
We need the following lemma to prove our results (see [16]).
Lemma 1
Let\(X\in \mathcal{L} ( \mathcal{H} ) \)be a positive operator, and let\(x\in \mathcal{H}\)be any unit vector. If\(r\geq 1\), then
and if\(0\leq r\leq 1\), then
Before giving our next result, we set \(\Vert X \Vert _{ \operatorname{ber}}:=\sup \{ \vert \langle X \widehat{k}_{\lambda },\widehat{k}_{\mu } \rangle \vert : \lambda ,\mu \in \varOmega \} \) and \(m ( X ) := \inf_{\lambda \in \varOmega } \vert \widetilde{X} ( \lambda ) \vert \).
Theorem 2
Let\(X,Y,Z\in \mathcal{L} ( \mathcal{H} ) \)be operators such thatX, Yare positive. If\(p\geq q>1\)with\(\frac{1}{p}+ \frac{1}{q}=1\), then
for all\(r\geq \frac{2}{q}\).
Proof
Using the Cauchy–Schwarz inequality, we get
for all \(\lambda \in \varOmega \). By using the Young inequality and (2), we get
and it follows from inequality (4) that
for all \(\lambda \in \varOmega \). Since \(\widetilde{ ( \frac{X^{\frac{rp}{2}}}{p}+\frac{ ( Z^{\star }YZ ) ^{\frac{rq}{2}}}{q} ) } ( \lambda ) \) is positive, then we have
for all \(\lambda \in \varOmega \). This implies that
□
Taking the \(Z=I\) in inequality (5), we have the following result.
Corollary 3
Let\(X,Y\in \mathcal{L} ( \mathcal{H} ) \)be positive operators, and let\(p\geq q>1\)with\(\frac{1}{p}+\frac{1}{q}=1\). Then
for all\(r\geq \frac{2}{q}\).
Corollary 4
Let\(X,Y\in \mathcal{L} ( \mathcal{H} ) \)be positive operators. Then
Proof
As in the same arguments in the proof of Theorem 2, if we put \(r=p=q=2\), then we get
Since \([ \widetilde{X} ( \lambda ) ] ^{2} \geq 0\), \([ \widetilde{Y} ( \lambda ) ] ^{2} \geq 0\), and \(\widetilde{ ( X^{2}+Y^{2} ) } ( \lambda ) \geq 0\), taking the supremum over \(\lambda \in \varOmega \), we get that
□
Proposition 5
Let\(X,Y,Z\in \mathcal{L} ( \mathcal{H} ) \)such thatX, Yare positive, and let\(\frac{1}{p}+\frac{1}{q}=1\). Then
for all\(r\geq \frac{2}{q}\).
Proof
Indeed, for every \(\lambda ,\mu \in \varOmega \), we have
so that if we take the supremum over \(\lambda ,\mu \in \varOmega \) in inequality (6), we get
□
Remark 6
It follows from inequality
and inequality (6) that
Proposition 7
Let\(X,Y, Z\in \mathcal{L} ( \mathcal{H} ) \)such thatX, Yare positive, and let\(p\geq q>1\), where\(\frac{1}{p}+ \frac{1}{q}=1\). Then
for all\(r\geq \frac{2}{q}\).
Proof
By inequality (2), we have
for all \(\lambda ,\mu \in \varOmega \) and taking supremum over \(\lambda ,\mu \in \varOmega \) in the above inequality, we get
□
Now, we present the next lemma to obtain our last results.
Lemma 8
([16])
If\(f, g: [0, \infty )\longrightarrow \mathcal{R} \)are nonnegative continuous such that\(f(t)g(t)=t\) (\(t\in [0, \infty )\)), then
where\(X \in {\mathcal{L}}({\mathcal{H}})\)and\(x, y\in {\mathcal{H}}\).
In the next theorem we show an upper bound for the generalized Euclidean Berezin number.
Theorem 9
Let\(X_{i}, Y_{i}, Z_{i}, \in {\mathcal{L}}({\mathcal{H}})\) (\(1 \leq i \leq n\)). Then
where\(f, g: [0, \infty )\longrightarrow \mathcal{R} \)are nonnegative continuous such that\(f(t)g(t)=t\) (\(t\in [0, \infty )\)) and\(p, r\geq 1\).
Proof
For any \(\hat{k}_{\lambda }\in {\mathcal{H}(\varOmega )}\), we have
By taking the supremum on \(\hat{k}_{\lambda }\in {\mathcal{H}}\) with \(\|\hat{k}_{\lambda }\|=1\), we reach the desired inequality. □
Selecting \(X_{i}=Y_{i}=I\) for \(i=1,2,\ldots ,n\) in Theorem 9, we get the next result.
Corollary 10
Let\(Z_{i}\in {\mathcal{L}}({\mathcal{H}}) \) (\(1\leq i \leq n\)) and\(r, p\geq 1\). Then
where\(f, g: [0, \infty )\longrightarrow \mathcal{R} \)are nonnegative continuous such that\(f(t)g(t)=t\) (\(t\in [0, \infty )\)).
In particular, if\(X, Y\in {\mathcal{L}}({\mathcal{H}})\), then for all\(p\geq 1\)and\(0\leq \nu \leq 1\)
where
In the last theorem, we show another upper bound for \(\operatorname{ber}_{p}(T_{1},\ldots,T_{n})\).
Theorem 11
Let\(Z_{i}\in {\mathcal{L}}({\mathcal{H}})\) (\(1\leq i \leq n\)). Then
where\(p\geq 1\), \(0\leq \nu \leq 1\), and\(\delta (\hat{k}_{\lambda })= (\sqrt{ \langle |Z_{i}^{*}|^{2(1-\nu )}\hat{k}_{\lambda }, \hat{k}_{\lambda }\rangle } - \sqrt{ \langle |Z_{i}|^{2\nu }\hat{k} _{\lambda }, \hat{k}_{\lambda }\rangle } )^{2} \).
Proof
Let \(\hat{k}_{\lambda }\in {\mathcal{H}(\varOmega )}\). Then, by using Lemma 8 and inequality (3), we have
Thus
If we get the supremum over all \(\hat{k}_{\lambda }\in {\mathcal{H}( \varOmega )}\) with \(\|\hat{k}_{\lambda }\|=1\), then we reach the desired result. □
References
Ando, T.: Matrix Young inequality. Oper. Theory, Adv. Appl. 75, 33–38 (1995)
Aronzajn, N.: Theory of reproducing kernels. Trans. Am. Math. Soc. 68, 337–404 (1950)
Bakherad, M.: Some Berezin number inequalities for operator matrices. Czechoslov. Math. J. 68(143)(4), 997–1009 (2018)
Bakherad, M., Garayev, M.T.: Berezin number inequalities for operators. Concr. Oper. 6, 33–43 (2019)
Bakherad, M., Lashkaripour, R., Hajmohamadi, M., Yamanci, U.: Complete refinements of the Berezin number inequalities. J. Math. Inequal. 13(4), 1117–1128 (2019)
Basaran, H., Gürdal, M., Güncan, A.N.: Some operator inequalities associated with Kantorovich and Holder–McCarthy inequalities and their applications. Turk. J. Math. 43(1), 523–532 (2019)
Berezin, F.A.: Covariant and contravariant symbols for operators. Math. USSR, Izv. 6, 1117–1151 (1972)
Bhatia, R., Kittaneh, F.: On the singular values of a product of operators. SIAM J. Matrix Anal. Appl. 11, 272–277 (1990)
Garayev, M.T., Gürdal, M., Okudan, A.: Hardy–Hilbert’s inequality and a power inequality for Berezin numbers for operators. Math. Inequal. Appl. 19(3), 883–891 (2016)
Garayev, M.T., Gürdal, M., Saltan, S.: Hardy type inequality for reproducing kernel Hilbert space operators and related problems. Positivity 21(4), 1615–1623 (2017)
Gustafson, K.E., Rao, D.K.M.: Numerical Range. Springer, New York (1997)
Hajmohamadi, M., Lashkaripour, R., Bakherad, M.: Improvements of Berezin number inequalities. Linear Multilinear Algebra (to appear). https://doi.org/10.1080/03081087.2018.1538310
Karaev, M.T.: Berezin symbol and invertibility of operators on the functional Hilbert spaces. J. Funct. Anal. 238, 181–192 (2006)
Karaev, M.T.: Reproducing kernels and Berezin symbols techniques in various questions of operator theory. Complex Anal. Oper. Theory 7, 983–1018 (2013)
Karaev, M.T., Gürdal, M., Yamanci, U.: Some results related with Berezin symbols and Toeplitz operators. Math. Inequal. Appl. 17(3), 1031–1045 (2014)
Kittaneh, F.: Notes on some inequalities for Hilbert space operators. Publ. Res. Inst. Math. Sci. 24, 283–293 (1988)
Kittaneh, F., Manasrah, Y.: Improved Young and Heinz inequalities for matrices. J. Math. Anal. Appl. 361, 262–269 (2010)
Saitoh, S., Sawano, Y.: Theory of Reproducing Kernels and Applications. Springer, Singapore (2016)
Salemi, A., Sheikh Hosseini, A.: Matrix Young numerical radius inequalities. J. Math. Inequal. 16, 783–791 (2013)
Salemi, A., Sheikh Hosseini, A.: On reversing of the modified Young inequality. Ann. Funct. Anal. 5, 69–75 (2014)
Yamancı, U., Gürdal, M.: On numerical radius and Berezin number inequalities for reproducing kernel Hilbert space. N.Y. J. Math. 23, 1531–1537 (2017)
Yamancı, U., Gürdal, M., Garayev, M.T.: Berezin number inequality for convex function in reproducing kernel Hilbert space. Filomat 31(18), 5711–5717 (2017)
Acknowledgements
The first author would like to thank the Tusi Mathematical Research Group (TMRG).
Availability of data and materials
Not applicable.
Funding
The authors received no financial support for this article.
Author information
Authors and Affiliations
Contributions
The authors contributed equally to the manuscript and read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare that they have no competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, 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 licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Bakherad, M., Yamancı, U. New estimations for the Berezin number inequality. J Inequal Appl 2020, 40 (2020). https://doi.org/10.1186/s13660-020-2307-0
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s13660-020-2307-0