A Fast Shift-Splitting Method for Singular Generalized Saddle Point Problems

Abstract

For the iteration solution of singular generalized saddle point problems, a fast shift-splitting iteration method based on shift-splitting technique and symmetric and skew-symmetric splitting with respect to the upper-left block of the system matrix is proposed in this paper. Semi-convergence of the proposed method is carefully studied for singular case, and the conditions guaranteeing the semi-convergence are derived. Numerical experiments of a class of linearized Navier-Stokes equations are implemented to demonstrate the feasibility and effectiveness of the proposed method.