Iterative algorithms for the minimum-norm solution and the least-squares solution of the linear matrix equations \(A_1XB_1 + C_1X^TD_1 = M_1, A_2XB_2 + C_2 X^TD_2 = M_2\) (Q426323)

The authors focus on the solution of the generalized Sylvester matrix equations. A conjugate gradient algorithm (CG) is proposed. When the matrix equations are consistent, by the CG algorithm, a solution can be obtained within finite steps in the absence of rounding-errors for any initial point. When the matrix equations are inconsistent, applying the CG algorithm to the normal equation, a least squares solution with minimum norm can be obtained.