Schur inequality for coneigenvalues and conjugate-normal matrices
One of the consequences of the classical Schur theorem on triangularization of unitary matrices is the well known lower estimation for the Euclidean norm of the imaginary \(n \times n\) matrix \(A\): \[ \| A \|_E^{2}=\sum_{i,j=1}^n{| a_{i,j}|}^2 \geq{\sum_{i=1}^n{| \lambda_{i}|}^2}, \tag{1} \] where \(\lambda_{1}, \lambda_{2},\dots,\lambda_{n}\) are the eigenvalues of the matrix \(A\). If the matrix \(A\) is a nilpotent one then the estimation (1) is trivial. It turns out that it is possible to find an inequality as much universal as a simple one like the estimation (1). The inequality conserves significance in the nilpotent case and has the form \[ \| A \|_E^{2} \geq {\sum_{i=1}^n}{| \mu_i |}^2 \] \[ \mu_i = \sqrt{\nu_i},\qquad \text{Re }\mu_i \geq 0, \quad i=1, 2, \dots, n, \] where \(\mu_i\) are pseudo-eigenvalues of the matrix \(A\) and \({\nu_i}\) are eigenvalues of the matrix \(A_R = A\overline{A}\) or \( A_L =\overline{A}A\) (\(\overline{A}\) is a conjugate matrix for the matrix \(A\)).
This page was built for publication: Schur inequality for coneigenvalues and conjugate-normal matrices
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1288041)
