Numerical stability of Lanczos methods



Cahill, Eamonn, Irving, Alan, Johnston, Christopher and Sexton, James
(2000) Numerical stability of Lanczos methods. Nuclear Physics B - Proceedings Supplements, 83-84. pp. 825-827.

[thumbnail of Requires a viewer, such as GSview] Postscript (Requires a viewer, such as GSview)
9909131v1.ps.gz - Unspecified

Download (42kB)
[thumbnail of 614.pdf] PDF
614.pdf - Unspecified

Download (81kB)

Abstract

The Lanczos algorithm for matrix tridiagonalisation suffers from strong numerical instability in finite precision arithmetic when applied to evaluate matrix eigenvalues. The mechanism by which this instability arises is well documented in the literature. A recent application of the Lanczos algorithm proposed by Bai, Fahey and Golub allows quadrature evaluation of inner products of the form $\psi^\dagger g(A) \psi$. We show that this quadrature evaluation is numerically stable and explain how the numerical errors which are such a fundamental element of the finite precision Lanczos tridiagonalisation procedure are automatically and exactly compensated in the Bai, Fahey and Golub algorithm. In the process, we shed new light on the mechanism by which roundoff error corrupts the Lanczos procedure.

Item Type: Article
Additional Information: LTH 457. arXiv Number: arXiv:hep-lat/9909131v1. Available online 23 July 2003. Issue: April 2000. Proceedings of the XVIIth International Symposium on Lattice Field Theory.
Uncontrolled Keywords: Lanczos algorithm, matrix tridiagonalisation, Bai, Fahey and Golub, Lanczos procedure
Subjects: ?? QC ??
Divisions: Faculty of Science and Engineering > School of Physical Sciences > Mathematical Sciences
Depositing User: Symplectic Admin
Date Deposited: 28 Nov 2008 10:46
Last Modified: 17 Dec 2022 01:14
DOI: 10.1016/s0920-5632(00)91816-4
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/614