| Large-scale sparse linear systems with saddle-point structure are widely derived from a variety of scientific computations and engineering problems,such as computational fluid dynamics,constrained optimization control,structural mechanics,linear programming,circuit design and so forth.The rapid solution of such problems is one of the hot issues in recent years.In this paper,the regularized idea is applied to the SIMPLE-like preconditioner for solving the nonsymmetric saddle point systems[Z.-Z.Liang,G.-F.Zhang,J.Comput.Appl.Math.,302(2016)211-223],and then a more efficient preconditioner is obtained.Furthermore,we extend the new preconditioner to generalized saddle point problems.The spectrum properties of the corresponding preconditioned matrices are studied in details.Furthermore,the degree of the minimal polynomial of the corresponding preconditioned matrices are also studied.Numerical experiments are presented to illustrate the effectiveness and feasibility of these new preconditioners. |
