A New Preconditioner for Solving the Double Saddle Point Problem
By introducing a suitable symmetric positive definite matrix(without parameters)into the sub blocks of the coefficient matrix of the double saddle point problem,the difficulty of parameter selection can be effectively avoided.Based on this idea,in this paper,a new iterative method and preconditioner are proposed to solve the double saddle point problem.The convergence conditions of the iterative method are given,and the spectral analysis of the preconditioned coefficient matrix is performed.Numerical experiments have verified the effectiveness of this preconditioner.
Double Saddle Point ProblemPreconditionerSpectral RadiusConvergence Rate
万方数据
维普
