scientific article; zbMATH DE number 2246263
For an \(n\times m\) complex matrix there are found simple and effective lower bounds for the spectral norm. Among them: \[ \| A\| \geqslant (\frac {| r_1| ^2+\dotsc +| r_n| ^2}n)^{1/2}, \] where \(r_1,\dots, r_n\) are the row sums of \(A\); \[ \| A\| \geqslant \frac 1{\sqrt 2} \bigl\{ \| \text{\textbf{a}}_i\| ^2+\| \text{\textbf{a}}_k\| ^2 +[(\| \text{\textbf{a}}_i\| ^2-\| \text{\textbf{a}}_k\| ^2)^2 +4| \text{\textbf{a}}^*_i \text{\textbf{a}}_k| ^2]^{1/2} \bigr\}^{1/2}, \] where \((\text{\textbf{a}}_1, \dotsc, \text{\textbf{a}}_n)\) is the partition of \(A\) into columns; \[ \| A\| \geqslant \frac {\| (\sum_j \text{\textbf{a}}_j \text{\textbf{a}}_j^*)\text{\textbf{r}}\| } { \| (\text{\textbf{a}}_1^*\text{\textbf{r}} \dots \text{\textbf{a}}_n^* \text{\textbf{r}})^T\| },\text{ where }\text{\textbf{r}}= (r_1,\dots, r_n)^T. \] These bounds are studied on hundreds of randomly chosen matrices; the bounds are proved to be suprisingly good. They all have complexity \(O(n^2)\), and the best \(O(n^2)\) bound which is presented here seems to be better than all \(O(n^2)\) bounds found in the literature.
- Lower bounds for the spectral radius of a matrix
- Lower bounds for the Wiener norm in \(\mathbb{Z}_p^d\)
- Variational analysis of the Crouzeix ratio
- Probabilistic upper bounds for the matrix two-norm
- Lower bounds for tau coefficients and operator norms using composite matrix norms
- Lower bounds to the spectral gap of Davies generators
- Spectral aspect subconvex bounds for \(U_{n+1} \times U_n\)
- Lower bounds for column matrix approximations
- Verified bounds for singular values, in particular for the spectral norm of a matrix and its inverse
- Obtaining bounds on the two norm of a matrix from the splitting lemma
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5717271)