scientific article; zbMATH DE number 3974169

From MaRDI portal
Publication:3740127





[For the entire collection see Zbl 0589.00015.] Some iterative methods based on decomposition in subspaces are used to compute the minimal eigenvalue of the generalized eigenvalue problem (1) \(Au=\lambda Su\) where A, S are symmetric positive definite matrices. On an example of a matrix A which is the finite difference discretization of the Laplace operator on \(\Omega \subset {\mathbb{R}}^ 2\) the domain decomposition method is investigated using the bisection method for outer iterations and e.g. the Lanczos method as inner iterative process. In case A is an M-matrix or \(A_{\lambda}=A-\lambda S\) is a 2-cyclic matrix, the block-relaxation methods are implemented. Finally the fictitious component method is suggested to solve the problem (1). Some theoretical estimates of the number of arithmetic operations are included.












This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3740127)