On estimating the separation of two regular matrix pairs (Q315702)

This is a generalization of the result given by \textit{S. Xu} [Linear Algebra Appl. 262, 67--82 (1997; Zbl 0880.15028)] where the separation of two matrices is analyzed, i.e., \(\mathrm{sep}[A,C]=\min_{\|X\|_F=1}\|AX-XC\|_F\) (\(\|\cdot\|_F\) refers to Frobenius norm). This paper is about the separation of two matrix pairs \((A,B)\) and \((C,D)\) with \(A,B\in\mathbb{R}^{n\times n}\) and \(C,D\in\mathbb{R}^{m\times m}\). The separation is now defined as \(\mathrm{dif}[(A,B),(C,D)]=\min_{\|(X,Y)\|_F=1}\|(AX-YC,BX-YD)\|_F\). Both upper and lower bounds are given in this paper which are based on the estimation of the smallest singular values of the structured matrix of Kronecker products \[ T = \begin{pmatrix} I_m\otimes A & C\otimes I_n\\I_m\otimes B&D\otimes I_n\end{pmatrix} \] and either the Schur or the Weierstrass canonical form of the four matrices \(A\), \(B\), \(C\), \(D\).