The Chambolle-Pock method also converges weakly with $0 < θ\le 1$ and $τσ\|L\|^{2} < 4θ(2-θ)/(1 - 2θ+ 9θ^{2} - 4θ^{3})$

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Upadhyaya, Manu
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911574434775040
author Upadhyaya, Manu
author_facet Upadhyaya, Manu
contents The Chambolle-Pock method, also known as the primal-dual hybrid gradient method, is a standard first-order algorithm for convex-concave saddle-point problems and composite convex optimization involving two proper, lower semicontinuous, convex functions and a bounded linear operator $L$. We study its convergence in real Hilbert spaces for step sizes $τ,σ>0$ and relaxation parameter $0<θ\le 1$. We prove that, if $τσ|L|^{2} \leq 4θ(2-θ)/(1 - 2θ+ 9θ^{2} - 4θ^{3})$, then the ergodic duality gap converges at rate $O(1/k)$, and that, when the inequality is strict, the primal-dual iterates converge weakly to a KKT point. In particular, this extends the weak-convergence theory to the previously unexplored regime $0<θ\le 1/2$. The proof is based on a Lyapunov function that remains uniformly valid over the entire interval $0<θ\le 1$.
format Preprint
id arxiv_https___arxiv_org_abs_2604_06423
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle The Chambolle-Pock method also converges weakly with $0 < θ\le 1$ and $τσ\|L\|^{2} < 4θ(2-θ)/(1 - 2θ+ 9θ^{2} - 4θ^{3})$
Upadhyaya, Manu
Optimization and Control
47J25, 49M29, 65K05, 90C25, 93D30
The Chambolle-Pock method, also known as the primal-dual hybrid gradient method, is a standard first-order algorithm for convex-concave saddle-point problems and composite convex optimization involving two proper, lower semicontinuous, convex functions and a bounded linear operator $L$. We study its convergence in real Hilbert spaces for step sizes $τ,σ>0$ and relaxation parameter $0<θ\le 1$. We prove that, if $τσ|L|^{2} \leq 4θ(2-θ)/(1 - 2θ+ 9θ^{2} - 4θ^{3})$, then the ergodic duality gap converges at rate $O(1/k)$, and that, when the inequality is strict, the primal-dual iterates converge weakly to a KKT point. In particular, this extends the weak-convergence theory to the previously unexplored regime $0<θ\le 1/2$. The proof is based on a Lyapunov function that remains uniformly valid over the entire interval $0<θ\le 1$.
title The Chambolle-Pock method also converges weakly with $0 < θ\le 1$ and $τσ\|L\|^{2} < 4θ(2-θ)/(1 - 2θ+ 9θ^{2} - 4θ^{3})$
topic Optimization and Control
47J25, 49M29, 65K05, 90C25, 93D30
url https://arxiv.org/abs/2604.06423