Saturday, November 1

Lanczos-Type Methods

10:00 AM-12:00 PM
Room: Ballroom 2

Among the iterative methods for solving large linear systems or eigenvalue problems with a sparse nonsymmetric matrix, those that are based on the Lanczos process feature low computational cost and low memory requirements.In the last ten years much progress has been made regarding new, faster converging algorithms, avoiding breakdowns by look-ahead, understanding other stability problems, and taking appropriate measure to avoid or reduce them. New related results for the symmetric case are also of high relevance. In this minisymposium, the speakers will present some of the most recent findings. The presentations should be of interest to a wide audience of people applying or investigating iterative methods.

Organizer: Martin H. Gutknecht
ETH-Zentrum, Switzerland

These abstracts are posted as PostScript files at

10:00 On the Role of the Left Starting Vector in the Nonsymmetric Lanczos Algorithm
Anne Greenbaum, Courant Institute of Mathematical Sciences, New York University
10:30 Look-Ahead Lanczos Revisited
Martin H. Gutknecht, Organizer; and Marlis Hochbruck, Universitat Tubingen, Germany
11:00 The Main Effects of Rounding Errors in Krylov Solvers for Symmetric Linear Systems
Gerard Sleijpen and Henk van der Vorst, University of Utrecht, The Netherlands; and Jan Modersitzki, Medical University of Lubeck, Germany
11:30 Residual Smoothing: Does it Help at All?
Martin H. Gutknecht, Organizer

LA97 Homepage | Program Updates| Registration | Hotel Information | Transportation | Program-at-a-Glance |

MMD, 5/14/97
tjf, 6/13/97
MMD, 6/25/97