Monotone and nonmonotone linearized block coordinate descent methods for nonsmooth composite optimization problems

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Nabou, Yassine, Bourkhissi, Lahcen El, Stich, Sebastian U., Valkonen, Tuomo
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908407872618496
author Nabou, Yassine
Bourkhissi, Lahcen El
Stich, Sebastian U.
Valkonen, Tuomo
author_facet Nabou, Yassine
Bourkhissi, Lahcen El
Stich, Sebastian U.
Valkonen, Tuomo
contents In this paper, we introduce both monotone and nonmonotone variants of LiBCoD, a \textbf{Li}nearized \textbf{B}lock \textbf{Co}ordinate \textbf{D}escent method for solving composite optimization problems. At each iteration, a random block is selected, and the smooth components of the objective are linearized along the chosen block in a Gauss-Newton approach. For the monotone variant, we establish a global sublinear convergence rate to a stationary point under the assumption of bounded iterates. For the nonmonotone variant, we derive a global sublinear convergence rate without requiring global Lipschitz continuity or bounded iterates. Preliminary numerical experiments highlight the promising performance of the proposed approach.
format Preprint
id arxiv_https___arxiv_org_abs_2506_12397
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Monotone and nonmonotone linearized block coordinate descent methods for nonsmooth composite optimization problems
Nabou, Yassine
Bourkhissi, Lahcen El
Stich, Sebastian U.
Valkonen, Tuomo
Optimization and Control
In this paper, we introduce both monotone and nonmonotone variants of LiBCoD, a \textbf{Li}nearized \textbf{B}lock \textbf{Co}ordinate \textbf{D}escent method for solving composite optimization problems. At each iteration, a random block is selected, and the smooth components of the objective are linearized along the chosen block in a Gauss-Newton approach. For the monotone variant, we establish a global sublinear convergence rate to a stationary point under the assumption of bounded iterates. For the nonmonotone variant, we derive a global sublinear convergence rate without requiring global Lipschitz continuity or bounded iterates. Preliminary numerical experiments highlight the promising performance of the proposed approach.
title Monotone and nonmonotone linearized block coordinate descent methods for nonsmooth composite optimization problems
topic Optimization and Control
url https://arxiv.org/abs/2506.12397