English

Extragradient Sliding for Composite Non-Monotone Variational Inequalities

Optimization and Control 2025-02-06 v1

Abstract

Variational inequalities offer a versatile and straightforward approach to analyzing a broad range of equilibrium problems in both theoretical and practical fields. In this paper, we consider a composite generally non-monotone variational inequality represented as a sum of LqL_q-Lipschitz monotone and LpL_p-Lipschitz generally non-monotone operators. We applied a special sliding version of the classical Extragradient method to this problem and obtain better convergence results. In particular, to achieve ε\varepsilon-accuracy of the solution, the oracle complexity of the non-monotone operator QQ for our algorithm is O(Lp2/ε2)O\left(L_p^2/\varepsilon^2\right) in contrast to the basic Extragradient algorithm with O((Lp+Lq)2/ε2)O\left((L_p+L_q)^2/\varepsilon^2\right). The results of numerical experiments confirm the theoretical findings and show the superiority of the proposed method.

Keywords

Cite

@article{arxiv.2403.14981,
  title  = {Extragradient Sliding for Composite Non-Monotone Variational Inequalities},
  author = {Roman Emelyanov and Andrey Tikhomirov and Aleksandr Beznosikov and Alexander Gasnikov},
  journal= {arXiv preprint arXiv:2403.14981},
  year   = {2025}
}

Comments

12 pages, 1 algorithm, 3 figures

R2 v1 2026-06-28T15:29:33.055Z