Wasim Audeh | arXiv Science

arXiv2026

Integral Representations and Sharp Numerical Radius Inequalities for Hilbert Space Operators

We study the numerical radius \(ω(T)\) of a bounded linear operator \(T\) on a complex Hilbert space. Write \(T=A+iB\), where \(A\) and \(B\) are the real and imaginary parts of \(T\). A central result is the exact formula \[ ω(T)=\fracπ{2}\sup_{\|x\|=1}\int_0^1 \left|\left\langle (A\cos πt+B\sin πt)x,x\right\rangle\right|\,dt . \] For every \(n\in\mathbb N\), we also prove the power form \[ ω^{2n}(T)=\frac{4^n}{\binom{2n}{n}} \sup_{\|x\|=1}\int_0^1 \left\langle (A\cos πt+B\sin πt)x,x\right\rangle^{2n}\,dt . \] This yields sharp upper bounds; for \(n=1\), together with the matching integral lower bound, it recovers Kittaneh's classical two-sided inequality \[ \frac14\|T^*T+TT^*\|\le ω^2(T)\le \frac12\|T^*T+TT^*\|. \] Furthermore, we obtain exact numerical radius formulas for some block operator matrices, including a \(3\times3\) nilpotent matrix and longer alternating paths. The factor \(π/2\) is best possible.

math.FA↗

arXiv2026

The \(L^2\)-Cartesian Numerical Radius of Hilbert Space Operators

Let \(\mathcal{H}\) be a complex Hilbert space, and let \(T=H+iK\in\mathcal{B}(\mathcal{H})\) be the Cartesian decomposition of \(T\), where \(H=H^{*}\) and \(K=K^{*}\). We introduce the \(L^2\)-Cartesian numerical radius \[ ω_2(T)=\left(\frac{1}{2π}\int_0^{2π}\|\cosθH+\sinθK\|^2\,dθ\right)^{1/2}. \] We study its basic properties and compare it with the classical numerical radius \(ω(T)\). In particular, we establish the sharp estimate \[ \frac{1}{\sqrt{2}}ω(T)\le ω_2(T)\le ω(T), \] and establish lower and upper bounds in terms of the Cartesian parts \(H,K\). As a product, we obtain \[ \frac{1}{4}\left\|T^{*}T+TT^{*}\right\|\le ω_2(T)^2\le ω(T)^2. \]

math.FA↗