Hauptseite > Publikationsdatenbank > Optimizing rational filters for interior eigenvalue solvers |
Conference Presentation (Invited) | FZJ-2019-06911 |
; ;
2019
Please use a persistent id in citations: http://hdl.handle.net/2128/23783
Abstract: Rational filter functions can be used to improve the convergence of the so-called contour-based eigensolvers, a popular family of algorithms for the solution of the interior eigenvalue problem. This talk provide an overview on how such filters can be optimized for performance and convergence based on a non-convex weighted Least-Square scheme.The net result is an almost parameter-free minimization framework with an enhanced usability and productivity which leads to rational functions outperforming state-of-art filters.
![]() |
The record appears in these collections: |