arXiv · 2611.22819
Anderson Acceleration of SCF Iteration and Newton's Method for Solving Nonlinear Eigenvector Problems
Abstract
The self-consistent field (SCF) iteration is a gateway method for solving nonlinear eigenvector problems, also known as nonlinear eigenvalue problems with eigenvector dependency (NEPv). However, the convergence of SCF can be slow in practice. In this paper, we study two local acceleration techniques for SCF applied to unitarily invariant NEPv: Anderson acceleration and Newton acceleration. A central difficulty in developing these techniques is the non-uniqueness of the eigenvector basis, which complicates direct application of standard Anderson acceleration and introduces singularity into the correction equation of the standard Newton formulation. For Anderson acceleration, we present three formulations based on unitarily invariant parameterization, basis alignment, and Grassmannian geometry, and establish regularity properties of the associated SCF maps. For Newton acceleration, we introduce basis-aligned updates and derive a correction equation whose operator is nonsingular under suitable assumptions, and prove local quadratic convergence. Numerical experiments on NEPv arising from computational physics and data analysis confirm the theoretical findings and demonstrate the effectiveness of the proposed acceleration techniques.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Ding Lu, Tom Werner, Zhaojun Bai. 2026-11-17. Anderson Acceleration of SCF Iteration and Newton's Method for Solving Nonlinear Eigenvector Problems. https://arxiv.science/abs/2611.22819
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite arXiv Science by Indiciae: BibTeX · RIS.