Splitting the Forward-Backward Algorithm: A Full Characterization

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Åkerman, Anton, Chenchene, Enis, Giselsson, Pontus, Naldi, Emanuele
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916691356680192
author Åkerman, Anton
Chenchene, Enis
Giselsson, Pontus
Naldi, Emanuele
author_facet Åkerman, Anton
Chenchene, Enis
Giselsson, Pontus
Naldi, Emanuele
contents We study frugal splitting algorithms with minimal lifting for solving monotone inclusion problems involving sums of maximal monotone and cocoercive operators. Building on a foundational result by Ryu, we fully characterize all methods that use only individual resolvent evaluations, direct evaluations of cocoercive operators, and minimal memory resources while ensuring convergence via averaged fixed-point iterations. We show that all such methods are captured by a unified framework, which includes known schemes and enables new ones with promising features. Systematic numerical experiments lead us to propose three design heuristics to achieve excellent performances in practice, yielding significant gains over existing methods.
format Preprint
id arxiv_https___arxiv_org_abs_2504_10999
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Splitting the Forward-Backward Algorithm: A Full Characterization
Åkerman, Anton
Chenchene, Enis
Giselsson, Pontus
Naldi, Emanuele
Optimization and Control
47N10, 47H05, 47H09, 65K10, 90C25
We study frugal splitting algorithms with minimal lifting for solving monotone inclusion problems involving sums of maximal monotone and cocoercive operators. Building on a foundational result by Ryu, we fully characterize all methods that use only individual resolvent evaluations, direct evaluations of cocoercive operators, and minimal memory resources while ensuring convergence via averaged fixed-point iterations. We show that all such methods are captured by a unified framework, which includes known schemes and enables new ones with promising features. Systematic numerical experiments lead us to propose three design heuristics to achieve excellent performances in practice, yielding significant gains over existing methods.
title Splitting the Forward-Backward Algorithm: A Full Characterization
topic Optimization and Control
47N10, 47H05, 47H09, 65K10, 90C25
url https://arxiv.org/abs/2504.10999