Sharp inequalities for the numerical radii of block operator matrices

DOI10.1007/S10476-019-0002-5MaRDI QIDQ2288342zbMATH OpenOpenAlexFDO

Authors Mohammad Sal Moslehian, Jamal Rooin, Masoomeh Ghaderi Aghideh

Publication date 17 January 2020

Published in Analysis Mathematica (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1810.13093

zbMATH Keywords

convexitymixed Cauchy-Schwarz inequalitynumerical radiuspolar decomposition

Mathematics Subject Classification ID

Operator matrices (47A08) Numerical range, numerical radius (47A12) Norms (inequalities, more than one norm, etc.) of linear operators (47A30) Linear operator inequalities (47A63) Operators on Hilbert spaces (general) (47B02)

Abstract: In this paper, we present several sharp upper bounds for the numerical radii of the diagonal and off-diagonal parts of the $2imes2$ block operator matrix $$. Among extensions of some results of Kittaneh et al., it is shown that if $$, and $f$ and $g$ are non-negative continuous functions on $[0,infty)$ such that $f(t)g(t)=t,,(tgeq0)$, then for all nonnegative nondecreasing convex functions $h$ on $[0,infty)$ , we obtain that �egin{align*}hleft(w^r(T) ight)leq maxleft(left|frac{1}{p}hleft(f^{pr}(left|A ight|) ight)+ frac{1}{q}hleft(g^{qr}(left|A^* ight|) ight) ight|, left|frac{1}{p}hleft(f^{pr}(left|D ight|) ight)+ frac{1}{q}hleft(g^{qr}(left|D^* ight|) ight) ight| ight), end{align*} where $p,q>1$ with $frac1p+frac1q=1$ and $rmin(p,q)geq2$.

Review text

Let \(A,B,C,D\) be bounded linear Hilbert space operators. The authors obtain sharp upper bounds for the numerical radii \(w(T)\) and \(w(S)\) of the block operator matrices \(T = \left(\begin{smallmatrix} A & O \\ O & D \end{smallmatrix}\right)\) and \(S = \left(\begin{smallmatrix} 0 & B \\ C & 0 \end{smallmatrix}\right)\). For example, it is shown that, if \(f(t),g(t)\) are continuous non-negative functions on \([0,\infty)\) with \(f(t)g(t)=t\), then, for any non-negative non-decreasing convex function \(h(t)\) on \([0,\infty)\), the following inequality holds: \[h(w(T))\le \frac{1}{2} \max \left(||h(f^{2} (|A|))+h(g^{2} (|A|))||,\, ||h(f^{2} (|A|))+h(g^{2} (|A|))||\right).\] A similar inequality is proved for the numerical radius \(w(S)\). The paper also contains other variations of these inequalities.

Recommendations

• Numerical radius inequalities for certain \(2\times 2\) operator matrices. II
• Numerical range estimation of block operator matrices
• Quadratic numerical radius inequalities of off-diagonal \(2 \times 2\) block operator matrices
• Numerical radius inequalities for certain 2 2 operator matrices
• Sharp operator mean inequalities of the numerical radii

Cited in

(14)

• On Berezin number inequalities for operator matrices
• The numerical radius and positivity of block matrices
• Inequalities for the block projection operators
• Upper and lower bounds for numerical radii of block shifts
• A numerical radius inequality involving the generalized Aluthge transform
• Equalities and inequalities for norms of block imaginary circulant operator matrices
• Numerical range estimation of block operator matrices
• Sharp operator mean inequalities of the numerical radii
• Generalizations of numerical radius inequalities related to block matrices
• Some general quadratic numerical radius inequalities for the off-diagonal parts of \(2 \times 2\) block operator matrices
• New unitary equivalences for some operator matrices with applications
• Unitary equivalences for k -circulant operator matrices with applications
• On the Berezin number of operator matrices
• Inequalities for operators and operator pairs in Hilbert spaces

This page was built for publication: Sharp inequalities for the numerical radii of block operator matrices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2288342)