New version of mirror prox for variational inequalities with adaptation to inexactness

DOI10.1007/978-3-030-38603-0_31zbMATH Open1477.90113arXiv1907.13455OpenAlexW3098876547MaRDI QIDQ1982926FDOQ1982926

Mohammad S. Alkousa, Evgeniya A. Vorontsova, F. S. Stonyakin

Publication date: 14 September 2021

Abstract: Some adaptive analogue of the Mirror Prox method for variational inequalities is proposed. In this work we consider the adaptation not only to the value of the Lipschitz constant, but also to the magnitude of the oracle error. This approach, in particular, allows us to prove a complexity near $Oleft(frac1varepsilonlo{g}_{2}frac1varepsilonight)$ for variational inequalities for a special class of monotone bounded operators. This estimate is optimal for variational inequalities with monotone Lipschitz-continuous operators. However, there exists some error, which may be insignificant. The results of experiments on the comparison of the proposed approach with some known analogues are presented. Also, we discuss the results of the experiments for matrix games in the case of using non-Euclidean proximal setup.

Full work available at URL: https://arxiv.org/abs/1907.13455

zbMATH Keywords

adaptationvariational inequalitymatrix gamebounded operatorinexactnessLipschitz-continuous operatormirror prox

Mathematics Subject Classification ID

Complementarity and equilibrium problems and variational inequalities (finite dimensions) (aspects of mathematical programming) (90C33)

Cited In (5)

• An adaptive two-stage proximal algorithm for equilibrium problems in Hadamard spaces
• Adaptive two-stage Bregman method for variational inequalities
• Adaptive extraproximal algorithm for the equilibrium problem in Hadamard spaces
• Accelerated and unaccelerated stochastic gradient descent in model generality
• A Novel Algorithm with Self-adaptive Technique for Solving Variational Inequalities in Banach Spaces

This page was built for publication: New version of mirror prox for variational inequalities with adaptation to inexactness

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1982926)