Least-squares solution for inverse eigenpair problem of nonnegative definite matrices (Q1591963)

Given the diagonal canonical form \(\Lambda=\text{diag}\{\lambda_1, \lambda_2,\dots, \lambda_n\}\) and the associated modal matrix \(X= \{x_1, x_2,\dots, x_n\}\) for a positive semidefinite (unsymmetric) matrix, in this paper existence conditions for the following problems are proved: (I) Find positive semidefinite (unsymmetric) matrices \(A\) such that \[ \|AX-XA \|= \min. \] (II) Given \(\widetilde A\in\mathbb{R}^{n \times n}\), find \(\widehat A\in S_E\) (solution set of (I)) such that \[ \|\widetilde A-\widehat A\|= \min_{A\in S_E} \|\widetilde A-A\|, \] where \(\|\cdot\|\) is the Frobenius norm. Solutions for (I) and (II) are derived and sufficient conditions that there exists an explicit solution for (I) and (II) are proved.