An iterative method for the symmetric and skew symmetric solutions of a linear matrix equation \(AXB+CYD=E\) (Q848561)

The computation of pairs of symmetric and skew symmetric solutions \(X\), \(Y\) of the linear matrix equation \(AXB + CYD = E\) with real matrices \(A\), \(B\), \(C\), \(D\), and \(E\) is considered. Iterative methods for computing such pairs of matrices are presented. It is shown that for any initial symmetric or skew symmetric pair of matrices \(X_0\), \(Y_0\) the algorithm gives a solution of the matrix equation by a finite number of iteration steps supposed the matrix equation is consistent. Using a special kind of the initial pair of matrices \(X_0\), \(Y_0\) the minimum norm solution can be obtained. Additionally, the following problems are considered: Let a pair of matrices \({\overline X}\), \({\overline Y}\) be given. Find the symmetric (skew symmetric) pair \({\widehat X}\), \({\widehat Y}\) of the solution set of the matrix equation such that \(\| {\widehat X} - {\overline X} \| + \| {\widehat Y} - {\overline Y} \| = \min \{ \| X - {\overline X} \| + \| Y - {\overline Y} \|\}\), where the minimum is taken over all pairs of matrices \(X\), \(Y\) of the solution set. The solution of these problems can be obtained by finding the minimum norm solution of the matrix equation \(A {\widetilde X}B + C {\widetilde Y}D = E - A{\overline X} B - C {\overline Y} D\).