An iterative method for solving a kind of constrained linear matrix equations system (Q966219)

The authors motivate their work by the noting that all the linear matrix equations (such as Lyapunov matrix equation, Sylvester equation, Stein equation and so on) can be rewritten as the following linear matrix equations system \[ \left[ \mathcal{A}_{1}(X),\mathcal{A}_{2}(X),\dots,\mathcal{A}_{r}(X)\right] = \left[ E_{1},E_{2},\dots,E_{r}\right], \] where \(\mathcal{A}_{i}\in LC\,^{m\times n,\,p_{i}\times q_{i}}\) and \(E_{i}\in C\,^{p_{i}\times q_{i}},\) \(i=1,2,\dots,r\) and that many common constraints are special cases of the following constraint \[ X=\mathcal{U}(X), \] where \(\mathcal{U\in }\) \(LC\,^{m\times n}\) is a selfconjugate involution operator. Then an iterative method is presented to solve the above linear matrix equations system with the above constraint and its associated optimal approximation problem. The algorithm is provided in details and allows to obtain a solution in a finite number of iteration steps, in the absence of roundoff errors, by any initial conditions \(X_{0}\). Moreover, when a special kind of initial matrix is chosen, the least Frobenius norm solution is obtained. Two numerical examples highlight the efficiency of the proposed iterative method.