Proximal Limited-Memory Quasi-Newton Methods for Nonsmooth Nonconvex Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dahl, Simeon vom, De Marchi, Alberto, Kanzow, Christian
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913115882389504
author Dahl, Simeon vom
De Marchi, Alberto
Kanzow, Christian
author_facet Dahl, Simeon vom
De Marchi, Alberto
Kanzow, Christian
contents We introduce a proximal limited--memory quasi--Newton scheme for minimizing the sum of a continuously differentiable function and a proper, lower semicontinuous and prox-bounded, possibly nonsmooth, function. Both functions might be nonconvex. The method builds upon the computation of scaled proximal operators and is globalized by adaptively updating a regularization parameter based on a criterion of sufficient decrease. We prove global convergence under mild assumptions and then establish convergence of the entire sequence (with rates) under the Kurdyka--Lojasiewicz property. To efficiently solve the subproblems, we exploit the compact representation of limited-memory quasi-Newton updates. We derive also a compact representation of the limited--memory Kleinmichel formula, a rank-one quasi-Newton scheme that preserves positive definiteness under the same condition as the BFGS update. Numerical results show a significant speed up compared to other methods.
format Preprint
id arxiv_https___arxiv_org_abs_2605_11627
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Proximal Limited-Memory Quasi-Newton Methods for Nonsmooth Nonconvex Optimization
Dahl, Simeon vom
De Marchi, Alberto
Kanzow, Christian
Optimization and Control
65K10, 90C06, 90C26, 90C53
We introduce a proximal limited--memory quasi--Newton scheme for minimizing the sum of a continuously differentiable function and a proper, lower semicontinuous and prox-bounded, possibly nonsmooth, function. Both functions might be nonconvex. The method builds upon the computation of scaled proximal operators and is globalized by adaptively updating a regularization parameter based on a criterion of sufficient decrease. We prove global convergence under mild assumptions and then establish convergence of the entire sequence (with rates) under the Kurdyka--Lojasiewicz property. To efficiently solve the subproblems, we exploit the compact representation of limited-memory quasi-Newton updates. We derive also a compact representation of the limited--memory Kleinmichel formula, a rank-one quasi-Newton scheme that preserves positive definiteness under the same condition as the BFGS update. Numerical results show a significant speed up compared to other methods.
title Proximal Limited-Memory Quasi-Newton Methods for Nonsmooth Nonconvex Optimization
topic Optimization and Control
65K10, 90C06, 90C26, 90C53
url https://arxiv.org/abs/2605.11627