Algebraic Multigrid Preconditioning for Iterative Eigensolvers

The paper presents a comparative study of iterative solvers for eigen- problems, which arise e.g. in solid mechanics or structural analysis. We consider problems obtained by discretization of elliptic and self-adjoint partial differential operators. Typically, only a few of the smallest eigen- values of these problems are to be computed. We discuss various gradient based preconditioned eigensolvers which make use of algebraic multigrid preconditioning. We present algorithms together with numerical results. Performance characteristics are derived by a comparison with the solution of test problems. We show that known advantages of algebraic multigrid preconditioning (e.g. for boundary-value problems with large jumps in the coefficients) transfer to the eigensolvers considered here.