Neural Network Acceleration of Iterative Methods for Nonlinear Schrödinger Eigenvalue Problems

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Peterseim, Daniel, Pietschmann, Jan-F., Püschel, Jonas, Ruess, Kilian
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908460782714880
author Peterseim, Daniel
Pietschmann, Jan-F.
Püschel, Jonas
Ruess, Kilian
author_facet Peterseim, Daniel
Pietschmann, Jan-F.
Püschel, Jonas
Ruess, Kilian
contents We present a novel approach to accelerate iterative methods to solve nonlinear Schrödinger eigenvalue problems using neural networks. Nonlinear eigenvector problems are fundamental in quantum mechanics and other fields, yet conventional solvers often suffer from slow convergence in extreme parameter regimes, as exemplified by the rotating Bose- Einstein condensate (BEC) problem. Our method uses a neural network to predict and refine solution trajectories, leveraging knowledge from previous simulations to improve convergence speed and accuracy. Numerical experiments demonstrate significant speed-up over classical solvers, highlighting both the strengths and limitations of the approach.
format Preprint
id arxiv_https___arxiv_org_abs_2507_16349
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Neural Network Acceleration of Iterative Methods for Nonlinear Schrödinger Eigenvalue Problems
Peterseim, Daniel
Pietschmann, Jan-F.
Püschel, Jonas
Ruess, Kilian
Numerical Analysis
We present a novel approach to accelerate iterative methods to solve nonlinear Schrödinger eigenvalue problems using neural networks. Nonlinear eigenvector problems are fundamental in quantum mechanics and other fields, yet conventional solvers often suffer from slow convergence in extreme parameter regimes, as exemplified by the rotating Bose- Einstein condensate (BEC) problem. Our method uses a neural network to predict and refine solution trajectories, leveraging knowledge from previous simulations to improve convergence speed and accuracy. Numerical experiments demonstrate significant speed-up over classical solvers, highlighting both the strengths and limitations of the approach.
title Neural Network Acceleration of Iterative Methods for Nonlinear Schrödinger Eigenvalue Problems
topic Numerical Analysis
url https://arxiv.org/abs/2507.16349