Searcharxiv⌕ Search

arXiv subjects

Pintu Bhunia

Publications and source records attributed to Pintu Bhunia.

At least 37 records · Page 2Linked to original sources

Power numerical radius inequalities from an extension of Buzano's inequality

Several numerical radius inequalities are studied by developing an extension of the Buzano's inequality. It is shown that if $T$ is a bounded linear operator on a complex Hilbert space, then \begin{eqnarray*} w^n(T) &\leq& \frac{1}{2^{n-1}} w(T^n)+ \sum_{k=1}^{n-1} \frac{1}{2^{k}} \left\|T^k \right\| \left\|T \right\|^{n-k}, \end{eqnarray*} for every positive integer $n\geq 2.$ This is a non-trivial improvement of the classical inequality $w(T)\leq \|T\|.$ The above inequality gives an estimation for the numerical radius of the nilpotent operators, i.e., if $T^n=0$ for some least positive integer $n\geq 2$, then \begin{eqnarray*} w(T) &\leq& \left(\sum_{k=1}^{n-1} \frac{1}{2^{k}} \left\|T^k \right\| \left\|T \right\|^{n-k}\right)^{1/n} \leq \left( 1- \frac{1}{2^{n-1}}\right)^{1/n} \|T\|. \end{eqnarray*} Also, we deduce a reverse inequality for the numerical radius power inequality $w(T^n)\leq w^n(T)$. We show that if $\|T\|\leq 1$, then \begin{eqnarray*} w^n(T) &\leq& \frac{1}{2^{n-1}} w(T^n)+ 1- \frac{1}{2^{n-1}}, \end{eqnarray*} for every positive integer $n\geq 2.$ This inequality is sharp.

math.FA↗

Euclidean operator radius inequalities of $d$-tuple operators and operator matrices

In this paper, we develop several Euclidean operator radius inequalities of $d$-tuple operators, as well as the sum and the product of $d$-tuple operators. Also, we obtain a power inequality for the Euclidean operator radius. Further, we develop Euclidean operator radius inequalities of $2\times 2$ operator matrices whose entries are $d$-tuple operators.

math.FA↗

Sharper bounds for the numerical radius of $ \lowercase{n}\times \lowercase{n}$ operator matrices

Let $A=\begin{bmatrix} A_{ij} \end{bmatrix}$ be an $n\times n$ operator matrix, where each $A_{ij}$ is a bounded linear operator on a complex Hilbert space. Among other numerical radius bounds, we show that $w(A)\leq w(\hat{A})$, where $\hat{A}=\begin{bmatrix} \hat{a}_{ij} \end{bmatrix}$ is an $n\times n$ complex matrix, with $$\hat{a}_{ij}= \begin{cases} w(A_{ii}) \text{ when $i=j$,} \left \| | A_{ij}|+ | A_{ji}^*| \right\|^{1/2} \left\| | A_{ji}|+ | A_{ij}^*| \right\|^{1/2} \text{ when $i j$} . \end{cases}$$ This is a considerable improvement of the existing bound $w(A)\leq w(\tilde{A})$, where $\tilde{A}=\begin{bmatrix} \tilde{a}_{ij} \end{bmatrix}$ is an $n\times n$ complex matrix, with $$\tilde{a}_{ij}= \begin{cases} w(A_{ii}) \text{ when $i=j$}, \|A_{ij}\| \text{ when $i\neq j$}. \end{cases}$$ Further, applying the bounds, we develop the numerical radius bounds for the product of two operators and the commutator of operators. Also, we develop an upper bound for the spectral radius of the sum of the product of $n$ pairs of operators, which improve the existing bound.

math.FA↗

Improved bounds for the numerical radius via polar decomposition of operators

Using the polar decomposition of a bounded linear operator $A$ defined on a complex Hilbert space, we obtain several numerical radius inequalities of the operator $A$, which generalize and improve the earlier related ones. Among other bounds, we show that if $w(A)$ is the numerical radius of $A$, then \begin{eqnarray*} w(A) &\leq& \frac12 \|A\|^{1/2} \left\| |A|^{t} + |A^*|^{1-t} \right \|, \end{eqnarray*} for all $t\in [0,1].$ Also, we obtain some upper bounds for the numerical radius involving the spectral radius and the Aluthge transform of operators. It is shown that \begin{eqnarray*} w(A) &\leq& \|A\|^{1/2} \left( \frac12 \left \| \frac{ |A|+|A^*|}2 \right\| +\frac12 \left\| \widetilde{A}\right \| \right)^{1/2}, \end{eqnarray*} where $\widetilde{A}= |A|^{1/2}U|A|^{1/2} $ is the Aluthge transform of $A$ and $A=U|A|$ is the polar decomposition of $A$. Other related results are also provided.

math.FA↗

Numerical radius inequalities of bounded linear operators and $(α,β)$-normal operators

We obtain various upper bounds for the numerical radius $w(T)$ of a bounded linear operator $T$ defined on a complex Hilbert space $\mathcal{H}$, by developing the upper bounds for the $α$-norm of $T$, which is defined as $\|T\|_α= \sup \left\{ \sqrt{α|\langle Tx,x \rangle|^2+ (1-α)\|Tx\|^2 } : x\in \mathcal{H}, \|x\|=1 \right\}$ for $ 0\leq α\leq 1 $. Further, we prove that \begin{eqnarray*} w(T) &\leq & \sqrt{\left( \min_{α\in [0,1]}\left\| α|T|+(1-α)|T^*| \right\| \right) \|T\|} \,\,\,\, \leq \,\, \,\, \|T\|. \end{eqnarray*} For $0\leq α\leq 1 \leq β,$ the operator $T$ is called $(α,β)$-normal if $α^2 T^*T\leq TT^*\leq β^2 T^*T$ holds. Note that every invertible operator is an $(α,β)$-normal operator for suitable values of $α$ and $β$. Among other lower bound for the numerical radius of an $(α,β)$-normal operator $T$, we show that \begin{eqnarray*} w(T) &\geq & \sqrt{\max \left\{ 1+α^2, 1+\frac{1}{β^2}\right\} \frac{\|T\|^2}{4}+ \frac {\left| \|\Re(T)\|^2-\|\Im(T)\|^2 \right|}2} &\geq & \max \left\{ \sqrt{1+α^2}, \sqrt{1+\frac{1}{β^2}} \right\} \frac{\|T\|}{2} & > & \frac{\|T\|}2, \end{eqnarray*} where $\Re(T)$ and $\Im(T)$ are the real part and imaginary part of $T$, respectively.

math.FA↗

Numerical radius inequalities and estimation of zeros of polynomials

Let $A$ be a bounded linear operator defined on a complex Hilbert space and let $|A|=(A^*A)^{1/2}$ be the positive square root of $A$. Among other refinements of the well known numerical radius inequality $w^2(A)\leq \frac12 \|A^*A+AA^*\|$, we show that \begin{eqnarray*} w^2(A)&\leq&\frac{1}{4} w^2 \left(|A|+i|A^*|\right)+\frac{1}{8}\left\||A|^2+|A^*|^2\right \|+\frac{1}{4}w\left(|A||A^*|\right) &\leq& \frac12 \|A^*A+AA^*\|. \end{eqnarray*} Also, we develop inequalities involving numerical radius and spectral radius for the sum of the product operators, from which we derive the following inequalities $$ w^p(A) \leq \frac{1}{\sqrt{2} } w(|A|^p+i|A^*|^p )\leq \|A\|^p$$ for all $p\geq 1.$ Further, we derive new bounds for the zeros of complex polynomials.

math.FA↗

Numerical radius inequalities of sectorial matrices

We obtain several upper and lower bounds for the numerical radius of sectorial matrices. We also develop several numerical radius inequalities of the sum, product and commutator of sectorial matrices. The inequalities obtained here are sharper than the existing related inequalities for general matrices. Among many other results we prove that if $A$ is an $n\times n$ complex matrix with the numerical range $W(A)$ satisfying $W(A)\subseteq\{re^{\pm iθ}~:~θ_1\leqθ\leqθ_2\},$ where $r>0$ and $θ_1,θ_2\in \left[0,π/2\right],$ then \begin{eqnarray*} &&(i)\,\, w(A) \geq \frac{cscγ}{2}\|A\| + \frac{cscγ}{2}\left| \|\Im(A)\|-\|\Re(A)\|\right|,\,\,\text{and} &&(ii)\,\, w^2(A) \geq \frac{csc^2γ}{4}\|AA^*+A^*A\| + \frac{csc^2γ}{2}\left| \|\Im(A)\|^2-\|\Re(A)\|^2\right|, \end{eqnarray*} where $γ=\max\{θ_2,π/2-θ_1\}$. We also prove that if $A,B$ are sectorial matrices with sectorial index $γ\in [0,π/2)$ and they are double commuting, then $w(AB)\leq \left(1+\sin^2γ\right)w(A)w(B).$

math.FA↗

Euclidean operator radius inequalities of a pair of bounded linear operators and their applications

We obtain several sharp lower and upper bounds for the Euclidean operator radius of a pair of bounded linear operators defined on a complex Hilbert space. As applications of these bounds we deduce a chain of new bounds for the classical numerical radius of a bounded linear operator which improve on the existing ones. In particular, we prove that for a bounded linear operator $A,$ \[\frac{1}{4} \|A^*A+AA^*\|+\fracμ{2}\max \{\|\Re(A)\|,\|\Im(A)\|\} \leq w^2(A) \, \leq \, w^2( |\Re(A)| +i |\Im(A)|),\] where $μ= \big| \|\Re(A)+\Im(A)\|-\|\Re(A)-\Im(A)\|\big|.$ This improve the existing upper and lower bounds of the numerical radius, namely, \[ \frac14 \|A^*A+AA^*\|\leq w^2(A) \leq \frac12 \|A^*A+AA^*\|. \]

math.FA↗

Numerical radius inequalities for tensor product of operators

The two well-known numerical radius inequalities for the tensor product $A \otimes B$ acting on $\mathbb{H} \otimes \mathbb{K}$, where $A$ and $B$ are bounded linear operators defined on complex Hilbert spaces $\mathbb{H} $ and $ \mathbb{K},$ respectively are, $ \frac{1}{2} \|A\|\|B\| \leq w(A \otimes B) \leq \|A\|\|B\| $ and $w(A)w(B) \leq w(A \otimes B) \leq \min \{ w(A) \|B\|, w(B) \|A\| \}. $ In this article we develop new lower and upper bounds for the numerical radius $w(A \otimes B)$ of the tensor product $A \otimes B $ and study the equality conditions for those bounds.

math.FA↗

Davis-Wielandt-Berezin radius inequalities of Reproducing kernel Hilbert space operators

Several upper and lower bounds of the Davis-Wielandt-Berezin radius of bounded linear operators defined on a reproducing kernel Hilbert space are given. Further, an inequality involving the Berezin number and the Davis-Wielandt-Berezin radius for the sum of two bounded linear operators is obtained, namely, if $A $ and $B$ are reproducing kernel Hilbert space operators, then $$η(A+B) \leq η(A)+η(B)+\textbf{ber}(A^*B+B^*A),$$ where $η(\cdot)$ and $\textbf{ber}(\cdot)$ are the Davis-Wielandt-Berezin radius and the Berezin number, respectively.

math.FA↗

Development of the Berezin number inequalities

We present new bounds for the Berezin number inequalities which improve on the existing bounds. We also obtain bounds for the Berezin norm of operators as well as the sum of two operators.

math.FA↗

Inequalities involving Berezin norm and Berezin number

We obtain new inequalities involving Berezin norm and Berezin number of bounded linear operators defined on a reproducing kernel Hilbert space $\mathscr{H}.$ Among many inequalities obtained here, it is shown that if $A$ is a positive bounded linear operator on $\mathscr{H}$, then $\|A\|_{ber}=\textbf{ber}(A)$, where $\|A\|_{ber}$ and $\textbf{ber}(A)$ are the Berezin norm and Berezin number of $A$, respectively. In contrast to the numerical radius, this equality does not hold for selfadjoint operators, which highlights the necessity of studying Berezin number inequalities independently.

math.FA↗

Improvement of numerical radius inequalities

We develop upper and lower bounds for the numerical radius of $2\times 2$ off-diagonal operator matrices, which generalize and improve on the existing ones. We also show that if $A$ is a bounded linear operator on a complex Hilbert space and $|A|$ stands for the positive square root of $A$, i.e., $|A|=(A^*A)^{1/2}$, then for all $r\geq 1$, $w^{2r}(A) \leq \frac{1}{4} \big \| |A|^{2r}+|A^*|^{2r} \big \| + \frac{1}{2} \min\left\{ \big \|\Re\big(|A|^r\, |A^*|^r \big) \big \|, w^r(A^2) \right\} $ where $w(A)$, $\|A\|$ and $\Re(A)$, respectively, stand for the numerical radius, the operator norm and the real part of $A$. This (for $r=1$) improves on existing well-known numerical radius inequalities.

math.FA↗

Improved inequalities for the numerical radius via Cartesian decomposition

We develop various lower bounds for the numerical radius $w(A)$ of a bounded linear operator $A$ defined on a complex Hilbert space, which improve the existing inequality $w^2(A)\geq \frac{1}{4}\|A^*A+AA^*\|$. In particular, for $r\geq 1$, we show that \begin{eqnarray*}\frac{1}{4}\|A^*A+AA^*\| \leq\frac{1}{2} \left( \frac{1}{2}\|\Re(A)+\Im(A)\|^{2r}+\frac{1}{2}\|\Re(A)-\Im(A)\|^{2r}\right)^{\frac{1}{r}} \leq w^{2}(A),\end{eqnarray*} where $\Re(A)$ and $\Im(A)$ are the real and imaginary parts of $A$, respectively. Furthermore, we obtain upper bounds for $w^2(A)$ refining the well-known upper bound $w^2(A)\leq \frac{1}{2} \left(w(A^2)+\|A\|^2\right)$. Separate complete characterizations for $w(A)=\frac{\|A\|}{2}$ and $w(A)=\frac{1}{2}\sqrt{\|A^*A+AA^*\|}$ are also given.

math.FA↗

Generalized $A$-numerical radius of operators and related inequalities

Let $A$ be a non-zero positive bounded linear operator on a complex Hilbert space $(\mathcal{H},\langle\cdot,\cdot\rangle)$. Let $ω_A(T)$ denote the $A$-numerical radius of an operator $T$ acting on the semi-Hilbert space $(\mathcal{H},\langle\cdot,\cdot\rangle_A)$, where $\langle x, y\rangle_{A} :=\langle Ax, y\rangle$ for all $x,y\in \mathcal{H}$. Let $N_A(\cdot)$ be a seminorm on the algebra of all $A$-bounded operators acting on $\mathcal{H}$ and let $T$ be an operator which admits $A$-adjoint. Then, we define the generalized $A$-numerical radius as $$ω_{N_A}(T)=\displaystyle{\sup_{θ\in \mathbb{R}}}\; N_A\left(\frac{e^{iθ}T+e^{-iθ}T^{\sharp_A}}{2}\right),$$ where $T^{\sharp_A}$ denotes a distinguished $A$-adjoint of $T$. We develop several generalized $A$-numerical radius inequalities from which follows the existing numerical radius and $A$-numerical radius inequalities. We also obtain bounds for generalized $A$-numerical radius of sum and product of operators. Finally, we study $ω_{N_A}(\cdot)$ in the setting of two particular seminorms $N_A(\cdot)$.

math.FA↗

New semi-norm of semi-Hilbertian space operators and its application

In this paper, we introduce a new semi-norm of operators on a semi-Hilbertian space, which generalizes the A-numerical radius and A-operator semi-norm. We study the basic properties of this semi-norm, including upper and lower bounds for it. As an application of this new semi-norm, we obtain upper bounds for the A-numerical radius of semi-Hilbertian space operators, which are sharper than the earlier ones.

math.FA↗

Annular bounds for the zeros of a polynomial from companion matrix

Let $p(z)=z^n+a_{n-1}z^{n-1}+a_{n-2}z^{n-2}+\ldots+a_1z+a_0$ be a complex polynomial with $a_0\neq 0$ and $n\geq 3$. Several new upper bounds for the moduli of the zeros of $p$ are developed. In particular, if $α=\sqrt{\sum_{j=0}^{n-1}|a_j|^2}$ and $z$ is any zero of $p$, then we show that \begin{eqnarray*} |z|^2 &\leq & \cos^2 \fracπ{n+1}+|a_{n-2}|+ \frac{1}{4} \left ( |a_{n-1}|+ { α} \right)^2 + \frac{1}{2}\sqrt{α^2-|a_{n-1}|^2} + \frac{1}{2}α, \end{eqnarray*} which is sharper than the Abu-Omar and Kittaneh's bound \begin{eqnarray*} |z|^2 &\leq & \cos^2 \fracπ{n+1}+ \frac{1}{4} \left ( |a_{n-1}|+ { α}\right)^2 + α \end{eqnarray*} if and only if $2|a_{n-2}|< \sqrt{\sum_{j=0}^{n-1}|a_j|^2}-\sqrt{\sum_{j=0}^{n-2}|a_j|^2}. $ The upper bounds obtained here enable us to describe smaller annuli in the complex plane containing all the zeros of $p$.

math.CV↗

$A$-numerical radius : New inequalities and characterization of equalities

We develope new lower bounds for the $A$-numerical radius of semi-Hilbertian space operators, and applying these bounds we obtain upper bounds for the $A$-numerical radius of the commutators of operators. The bounds obtained here improve on the existing ones. Further, we provide characterizations for the equality of the existing $A$-numerical radius inequalities of semi-Hilbertian space operators.

math.FA↗