Augmented Block Householder Arnoldi Method

Document Type

Article

Date of Original Version

11-1-2008

Abstract

Computing the eigenvalues and eigenvectors of a large sparse nonsymmetric matrix arises in many applications and can be a very computationally challenging problem. In this paper we propose the Augmented Block Householder Arnoldi (ABHA) method that combines the advantages of a block routine with an augmented Krylov routine. A public domain MATLAB code ahbeigs has been developed and numerical experiments indicate that the code is competitive with other publicly available codes. © 2008 Elsevier Inc. All rights reserved.

Publication Title, e.g., Journal

Linear Algebra and Its Applications

Volume

429

Issue

10

Share

COinS