SOR for \(AX-XB=C\) (Q1176522)

The authors consider a new approach to the block SOR method applied to linear systems of equations which can be written as a matrix equation \(AX-XB=C\). Such systems arise from finite differencing of separable elliptic boundary value problems on rectangular domains. The translation to the matrix equation enables us to handle an iterative method for the solution of such equations including Lyapunov's matrix equation as a special case. Further it gives us a more compact form of the problem of choosing optimal parameters for the block SOR method. This is particularly helpful for non-selfadjoint problems of elliptic type. Applying the technique, the optimal parameters for the model problem of a convection-diffusion equation are determined under more general assumptions than those of \textit{R. C. Y. Chin} and {T. A. Manteuffel} [SIAM J. Numer. Anal. 25, No. 3, 564-585 (1988; Zbl 0655.65060)]. Numerical data are given for the optimal parameters and the spectral radii of the optimized iteration matrices.