Monday Afternoon, October 23
MS8
Algorithms and Software for Large Scale Nonsymmetric Eigenvalue Problems
The last few years has seen a vigorous research effort in numerical methods
for large scale nonsymmetric eigenvalue problems. This effort is starting to be
realized in high quality software. The first aim of this minisymposium is to
review the current software and the algorithms. The software reviewed includes
a block nonsymmetric Lanczos algorithm, and the various Arnoldi algorithms
including the block, implicitly re-started and rational Krylov variants. The
second goal of the minisymposium is to discuss whether the software is solving
the large scale eigenvalue problems arising in industry.
Organizer: Richard B. Lehoucq
Rice University
- 2:00 Industrial Solution of Large Scale Unsymmetric Eigenvalue Problems with MSC/NASTRAN.
- Louis Komzsik, The MacNeal-Schwendler Corporation
- 2:30 ABLE Algorithm Template for Solving Large Scale Non-Hermitian Eigenvalue Problems
- Zhaojun Bai and D. Day, University of Kentucky; and Q. Ye, University of Manitoba, Canada
- 3:00 Predicting Singularities with the Rational Krylov Algorithm
- Axel Ruhe, Chalmers University of Technology, Sweden
- 3:30 An Evaluation of Software for Solving Large Sparse Unsymmetric Eigenvalue Problems
- Richard B. Lehoucq, Organizer, and J.A. Scott, Rutherford Appleton Laboratory, United Kingdom
7/25/95