The general coupled linear matrix equations with conjugate and transpose unknowns over the mixed groups of generalized reflexive and anti-reflexive matrices (Q2514086)

This paper studies solution sets \(\{X_i\}\) and\( \{Z_i\}\) for coupled linear matrix equations with \((P_i,Q_i)\) reflective or antireflective matrices \(X_i\) and \(Z_i\). Algorithm (1) establishes whether the linear matrix equation is solvable for such \(X_i\) and \(Z_i\) by using a conjugate gradient like 2-fold matrix iteration with mutually perpendicular iterates. Thus -- at least in theory -- the algorithm will converge in finitely many steps and it does so a bit erratically and more slowly than expected in two computed 5 by 5 and 2 by 2 matrix examples. If such a system is solvable, then algorithm (2) computes the optimal \((P_i,Q_i)\) reflective or antireflective matrix solution sets \(\{X_i\}\) and \(\{Z_i\}\). An extended description of earlier efforts in this realm is given, but the similar two stage linear algorithm of \textit{J. Huang} and \textit{L. Nong} [Linear Algebra Appl. 432, No. 5, 1176--1188 (2010; Zbl 1194.65053)] is omitted.