Click image for larger photo.
The aim of this workshop is to bring together leading researchers in the theory and numerical solution of eigenvalue problems with a view to surveying the state of the art, promoting collaboration, and making progress on the many challenging problems in this area. We will also invite leading users of these algorithms, especially those that strain the envelope of what can be done now, to present their work. A unique feature of the workshop will be the chance for researchers from all parts of the spectrum from core linear algebra to numerical linear algebra to interact and work together intensively for the duration of the workshop.
Specific objectives are
- To survey and discuss recent developments in both theoretical and computational aspects of matrix eigenvalue problems.
- To report and advance progress on challenging open problems in the field, including those arising in important practical applications.
- To identify new avenues of research on which progress could be made by combining theoretical and numerical expertise.
- To serve as a meeting place for initiating new collaborations, in particular between junior and more established researchers.
As an example of the sort of impact we hope the workshop will have, at the 1990 Householder Symposium in Tylosand, Sweden (a much larger meeting, but one with broadly similar objectives to the proposed workshop) Henk Van der Vorst presented his fledgling Bi-CGSTAB algorithm for solving nonsymmetric linear systems. The resulting paper, subsequently published in the SIAM Journal on Scientific Computing, was recently identified as the most heavily cited journal paper across the whole of mathematics published in the 1990s (see SIAM News, March 2001).
An important feature of the workshop is its focussed nature. All other conferences in core and numerical linear algebra are much broader, and so do not allow the focus on eigenvalue problems or the bringing together of a critical mass of leading eigenvalue researchers in an environment in which they can work together. (There is one *more* specialized meeting, the approximately biennial series of International Workshops on Accurate Solution of Eigenvalue Problems, which is much more numerically oriented than the proposed workshop and which has grown in size to be more like a typical conference.)
Click image for larger photo.