Norm Inequalities for Hilbert space operators with Applications

Pintu Bhunia

2024-04-02

Abstract:Several unitarily invariant norm inequalities and numerical radius inequalities for Hilbert space operators are studied. We investigate some necessary and sufficient conditions for the parallelism of two bounded operators. For a finite rank operator $A,$ it is shown that \begin{eqnarray*} \|A\|_{p} &\leq &\left(\textit{rank} \, A\right)^{1/{2p}} \|A\|_{2p} \,\, \leq \,\, \left(\textit{rank} \, A\right)^{{(2p-1)}/{2p^2}} \|A\|_{2p^2}, \quad \textit{for all $p\geq 1 $} \end{eqnarray*} where $\|\cdot\|_p$ is the Schatten $p$-norm. If $\{ \lambda_n(A) \}$ is a listing of all non-zero eigenvalues (with multiplicity) of a compact operator $A$, then we show that \begin{eqnarray*} \sum_{n} \left|\lambda_n(A)\right|^{p} &\leq& \frac12 \| A\|_{ p}^{ p} + \frac12 \| A^2\|_{p/2}^{p/2}, \quad \textit{for all $p\geq 2$} \end{eqnarray*} which improves the classical Weyl's inequality $\sum_{n} \left|\lambda_n(A)\right|^{p} \leq \| A\|_{ p}^{ p}$ [Proc. Nat. Acad. Sci. USA 1949]. For an $n\times n$ matrix $A$, we show that the function $p\to n^{-{1}/{p}}\|A\|_p$ is monotone increasing on $p\geq 1,$ complementing the well known decreasing nature of $p\to \|A\|_p.$ \indent As an application of these inequalities, we provide an upper bound for the sum of the absolute values of the zeros of a complex polynomial. As another application we provide a refined upper bound for the energy of a graph $G$, namely, $\mathcal{E}(G) \leq \sqrt{2m\left(\textit{rank Adj(G)} \right)},$ where $m$ is the number of edges, improving on a bound by McClelland in $1971$.

Functional Analysis

What problem does this paper attempt to address?

The problems that this paper attempts to solve mainly focus on the norm inequalities of Hilbert - space operators and their applications. Specifically, the author has studied several unitarily invariant norm inequalities and numerical radius inequalities related to Hilbert - space operators. These studies include but are not limited to the following aspects:

1. **Conditions for parallel operators**: The author has explored the necessary and sufficient conditions for two bounded operators to be parallel. Here, being parallel means that there exists a complex number \(\lambda\) with modulus 1 such that \(\|A+\lambda B\| = \|A\|+\|B\|\).

2. **Schatten \(p\)-norm inequalities for finite - rank operators**: For a finite - rank operator \(A\), the author has proved the following inequality:
\[
\|A\|_p\leqslant(\text{rank }A)^{1 / 2p}\|A\|_{2p}\leqslant(\text{rank }A)^{(2p - 1)/2p^{2}}\|A\|_{2p^{2}}, \quad\forall p\geqslant1
\]

3. **Eigenvalue inequalities for compact operators**: For the non - zero eigenvalues \(\{\lambda_n(A)\}\) of a compact operator \(A\), the author has improved the classical Weyl inequality and proved that:
\[
\sum_n|\lambda_n(A)|^p\leqslant\frac{1}{2}\|A\|_p^p+\frac{1}{2}\|A^2\|_{p/2}^{p/2}, \quad\forall p\geqslant2
\]

4. **Monotonicity of matrix functions**: For an \(n\times n\) matrix \(A\), the author has proved that the function \(p\rightarrow n^{- 1/p}\|A\|_p\) is monotonically increasing for \(p\geqslant1\).

5. **Upper bounds for polynomial zeros and graph energy**: As a result of the above inequalities, the author has provided an upper bound for the sum of the absolute values of the zeros of a complex polynomial and has given an improved upper bound for the energy of a graph:
\[
E(G)\leqslant\sqrt{2m(\text{rank }\text{Adj}(G))}
\]
where \(m\) is the number of edges of the graph and \(\text{Adj}(G)\) is the adjacency matrix of the graph.

Through these studies, the author has not only developed new inequalities but also improved some existing results, providing a theoretical basis for further research in related fields.