Generalized AOR method for the augmented system

Abstract

In this article, the accelerated over-relaxation (AOR) method is generalized for solving the saddle point problem or the augmented system. The successive over-relaxation (SOR)-like method is the special case of the generalized AOR (GAOR) method. The connection between the parameters and the eigenvalues of the iteration matrix of the GAOR method is given. Therefore, a necessary and sufficient condition for the convergence of the GAOR method is derived. Numerical examples are also given to show that the GAOR method is better than the SOR-like method in certain cases.