Wednesday, October 29

MS4
Recent Developments in Eigenvalue Perturbation Theory and Algorithms (Part I of II)

3:00 PM-5:00 PM
Room: Ballroom 1

Eigenvalues and eigenvectors are one of the central concerns in numerical linear algebra. There has been a lot of recent work in eigenvalue perturbation theory and accurate algorithms. A common thread linking much of this work is the realization that it is possible to obtain much more accurate results and tighter bounds that was thought possible. The speakers in this minisymposium will discuss perturbation theory for eigenvalues and eigenvectors -- in particular techniques for obtaining stronger and more precise perturbation bounds and how these improved perturbation bounds can be exploited to derive more accurate and faster algorithms for computing eigenvalues and eigenvectors. They will consider tridiagonal, full, structured and unstructured, symmetric and non-symmetric matrices.

Organizer: Roy Mathias
College of William & Mary

3:00 Spectral Variation for Diagonalizable Matrices
Ren-Cang Li, University of Kentucky; Rajendra Bhatia, Indian Statistical Institute, India; and Fuad Kittaneh, University of Jordan, Jordan
3:30 On the Lidskii-Mirdky-Wielandt Theorem
Chi-Kwong Li, College of William & Mary; and Roy Mathias, Organizer
4:00 First Order Eigenvalue Perturbation Theory and the Newton Diagram
Julio Moro, Universidad Carlos III, Spain; James V. Burke, University of Washington; and Michael L. Overton, New York University
4:30 When Are Factors of Indefinite Matrices Relatively Robust?
Inderjit Dhillon and Beresford Parlett, University of California, Berkeley

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


MMD, 4/15/97
tjf, 6/13/97
MMD, 10/7/97