A stochastic primal–dual algorithm for composite optimization with a linear operator

This paper introduces a stochastic primal–dual algorithm tailored for solving optimization problems involving the sum of three functions with a linear operator. Additionally, we conduct a comprehensive analysis of the convergence of our proposed algorithm within a generally convex framework. Our stu...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	Expert systems with applications Jg. 267; S. 126021
Hauptverfasser:	Wen, Meng, Zhang, Yongqiang, Tang, Yuchao, Cui, Angang, Peng, Jigen
Format:	Journal Article
Sprache:	Englisch
Veröffentlicht:	Elsevier Ltd 01.04.2025
Schlagworte:	Compositely optimization Primal–dual algorithm Stochastic algorithm Three-operator splitting 90C25 Three-operator splitting 49M29 65K10 Stochastic algorithm 49M27 47H05 Compositely optimization Primal–dual algorithm
ISSN:	0957-4174
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Beschreibung
Zusammenfassung:	This paper introduces a stochastic primal–dual algorithm tailored for solving optimization problems involving the sum of three functions with a linear operator. Additionally, we conduct a comprehensive analysis of the convergence of our proposed algorithm within a generally convex framework. Our study includes numerical experiments focusing on fused logistic regression and graph-guided regularized logistic regression problems. The results demonstrate that our algorithm outperforms other state-of-the-art methods in terms of efficiency and consistency. •Design a new stochastic primal–dual algorithm for solving Non-smooth optimization problems.•Conduct a comprehensive analysis of the convergence of our proposed algorithm.•Illustrate the efficiency of our proposed algorithm through some numerical example.
ISSN:	0957-4174
DOI:	10.1016/j.eswa.2024.126021