Substructuring Preconditioners for the Systems Arising from Plane Wave Discretization of Helmholtz Equations

Abstract

In this paper we are concerned with the plane wave methods for Helmholtz equations with large wave numbers. We extend the plane wave weighted least-squares (PWLS) method and the plane wave discontinuous Galerkin (PWDG) method to Helmholtz equations in inhomogeneous mediums and with mixed boundary conditions. The main goal of this paper is to construct parallel preconditioners for solving the resulting discrete systems. To this end, we design a kind of substructuring domain decomposition preconditioner for both the two-dimensional case and the three-dimensional case, which is different from the existing substructuring preconditioners. We apply the proposed preconditioner to solve the Helmholtz systems generated by the PWLS method or the PWDG method, and we find that the new preconditioners with practical coarse mesh sizes possess relatively stable convergence, i.e., the iteration counts of the corresponding iterative methods (PCG or preconditioned GMRES) increase slowly when the wave number increases (and the mesh size decreases).