Title:Variants of the PSS preconditioner for generalized saddle point problems from the Navier-Stokes equations

Abstract:The paper is concerned with the solution of generalized saddle point problems from the Navier-Stokes equations. A new preconditioner based on matrix splitting is presented, which can be viewed as an improved variant of the positive semidefinite and skew-Hermitian splitting preconditioner. Spectral analysis of the preconditioned coefficient matrix shows that it has an eigenvalue at 1 with high multiplicity and the other eigenvalues are contained in the unit disk centered at (1,0). The selection of optimal parameters is discussed, which improves the clustering of the eigenvalues of the preconditioned matrix. Moreover, we also provide a preconditioning variant of this preconditioner, which results in much better convergence properties and spectrum distributions. Numerical experiments are used to illustrate the advantages of the proposed preconditioners over some other preconditioners.