Global SCD algorithm for real positive definite linear systems with multiple right-hand sides (Q2381292)

The semi-conjugate direction method was already presented in literature for real positive definite systems by \textit{J. Y. Yuan, G. H. Golub, R. J. Plemmons}, and \textit{W. A. G. Cecilio} [BIT 44, No.~1, 189--207 (2004; Zbl 1052.65026)]. Starting from that, the authors introduce a global semi-conjugate direction method (Gl-SCD) for nonsymmetric linear systems with the same coefficient matrix \(\;A\) and different right-hand sides. Whenever \(A\) is symmetric positive definite, the method reduces to the usual conjugate gradient-type method. Firstly, left and right conjugate matrices are presented and their properties are established. Then, a complete left conjugate direction method is given together with a formal algorithm. The right conjugate direction method is not presented since its properties are quite the same. Lastly, Gl-SCD method is applied for solving the matrix Lyapunov equation and a nice numerical example in an efficient way.