Événement

Riemannian optimization methods for nonlinear eigenvector problems

UniDistance Suisse

Je 01.12.2022, 17:00 - 18:00

In this talk, we address the numerical solution of nonlinear eigenvector problems arising in computational physics and chemistry. These problems characterize critical points of the underlying energy function on the infinite-dimensional Stiefel manifold. To efficiently compute energy minimizers, we propose a novel Riemannian gradient descent method induced by an energy-adaptive metric.  The non-monotone line search and the inexact evaluation of Riemannian gradients substantially improve the overall efficiency of the method. Numerical experiments illustrate the performance of the method and demonstrates its competitiveness with well-established schemes.
(Joint work with R. Altmann and D. Peterseim)

Lien vers le site web: https://unidistance.ch/en/event/riemannian-optimization-methods-for-nonlinear-eigenvector-problems

Plus d'informations

Intervenants

  • University of Augsburg
    Tatjana Stykel
    Professor in the Institute of Mathematics and in the Centre for Advanced Analytics and Predictive Sciences (CAAPS)

Organisateur

Domaine

Sciences appliquées, technologie
Autres domaines

Type d'événement: Colloque/symposium/congrès

Public cible: Professionnel, Etudiant, Ecole - Secondaire II

Réservation

Sur réservation/inscription

Lieu de l'événement

L'événement se déroule en ligne.

Vers l'événement en ligne

géré par
Avec le soutien de