Saved in:
Bibliographic Details
Main Authors: Wei, Yi, Cai, Xufeng, Diakonikolas, Jelena
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.29128
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911556330061824
author Wei, Yi
Cai, Xufeng
Diakonikolas, Jelena
author_facet Wei, Yi
Cai, Xufeng
Diakonikolas, Jelena
contents Cyclic block coordinate methods are a fundamental class of first-order algorithms, widely used in practice for their simplicity and strong empirical performance. Yet, their theoretical behavior remains challenging to explain, and setting their step sizes -- beyond classical coordinate descent for minimization -- typically requires careful tuning or line-search machinery. In this work, we develop $\texttt{ADUCA}$ (Adaptive Delayed-Update Cyclic Algorithm), a cyclic algorithm addressing a broad class of Minty variational inequalities with monotone Lipschitz operators. $\texttt{ADUCA}$ is parameter-free: it requires no global or block-wise Lipschitz constants and uses no per-epoch line search, except at initialization. A key feature of the algorithm is using operator information delayed by a full cycle, which makes the algorithm compatible with parallel and distributed implementations, and attractive due to weakened synchronization requirements across blocks. We prove that $\texttt{ADUCA}$ attains (near) optimal global oracle complexity as a function of target error $ε>0,$ scaling with $1/ε$ for monotone operators, or with $\log^2(1/ε)$ for operators that are strongly monotone.
format Preprint
id arxiv_https___arxiv_org_abs_2603_29128
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Adaptive Delayed-Update Cyclic Algorithm for Variational Inequalities
Wei, Yi
Cai, Xufeng
Diakonikolas, Jelena
Optimization and Control
Machine Learning
Cyclic block coordinate methods are a fundamental class of first-order algorithms, widely used in practice for their simplicity and strong empirical performance. Yet, their theoretical behavior remains challenging to explain, and setting their step sizes -- beyond classical coordinate descent for minimization -- typically requires careful tuning or line-search machinery. In this work, we develop $\texttt{ADUCA}$ (Adaptive Delayed-Update Cyclic Algorithm), a cyclic algorithm addressing a broad class of Minty variational inequalities with monotone Lipschitz operators. $\texttt{ADUCA}$ is parameter-free: it requires no global or block-wise Lipschitz constants and uses no per-epoch line search, except at initialization. A key feature of the algorithm is using operator information delayed by a full cycle, which makes the algorithm compatible with parallel and distributed implementations, and attractive due to weakened synchronization requirements across blocks. We prove that $\texttt{ADUCA}$ attains (near) optimal global oracle complexity as a function of target error $ε>0,$ scaling with $1/ε$ for monotone operators, or with $\log^2(1/ε)$ for operators that are strongly monotone.
title Adaptive Delayed-Update Cyclic Algorithm for Variational Inequalities
topic Optimization and Control
Machine Learning
url https://arxiv.org/abs/2603.29128