scientific article; zbMATH DE number 26912

From MaRDI portal
Publication:3983548





The author [J. Complexity 2, 131-162 (1986; Zbl 0599.65021)] introduced the generalized minimal algorithm (gmr) for finding an approximation \((x,\rho)\) of a normalized eigenpair of a large symmetric matrix \(A\) in a Krylov space \({\mathcal K}_ 0\) which minimizes the Euclidean norm of the residual \(| Ax-\rho x|\) for \(| x|=1\), \(x\in{\mathcal K}_ 0\). In this paper, he describes an efficient implementation of his algorithm which is based on Golub's theorem on eigenvalues of a matrix which is perturbed by a rank one matrix. Extensive numerical experiments are reported where the residuals obtained by gmr algorithm are compared with those of the Lanczos algorithm. The method may be modified in order to find approximation to the extreme eigenpairs which minimize the residuals in \({\mathcal K}_ 0\).











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 Q3983548)