|
We study some iterative methods (such as the Lanczos method, GMRES and their other variants) and domain decomposition methods for solving nonlinear elliptic eigenvalue problems. First, we apply the Lanczos method to solve linear systems with multiple right hand side. We give an error bound for the approximate solution of the second linear system, where the Lanczos-Galerkin process is used to solve the first linear system. We also seek the possible application of the proposed numerical method to continuation problems. Next, we show some convergence theory for the MINRES, a variant of the Lanczos method. A multigrid-Lanczos and a multigrid-GMRES algorithm are proposed for tracking solution branches of associated discrete problem and detecting singular points along solution branches. The proposed algorithms have the advantage of being robust and easy to implement. Finally, we show how nonoverlapping and overlapping domain decomposition methods can be used to solve fourth order nonlinear elliptic eigenvalue problems. For the linearized von Kármán equation, we present preconditioners using both Fourier analysis and probing techniques for the interface systems, which are similar to those derived by Chan et al. Our numerical results show the efficiency of these algorithms.
|