Iterative solvers for BEM algebraic systems of equations (Q1961500)

In the first part of the paper, the authors give an overview over standard Krylov subspace iterative methods such as the conjugate gradient (CG) method for symmetric and positive definite systems and its generalization to unsymmetric systems of algebraic equations. The question of preconditioning is briefly discussed for the case of unsymmetric systems. The application of CG-like iterative methods to the solution of unsymmetric systems with dense and sometimes ill-conditioned system matrices usually arising from the boundary element collocation discretization of some boundary integral formulation of elliptic second-order boundary value problems is considered in the second part of the paper. The authors compare the convergence behaviour of several well-known CG-like methods such as CGS, Bi-CG, and Bi-CGStab with and without (simple) preconditioners. They consider 2D as well as 3D test problems.