301. A Fast Iterative Method for Variational Inequality and Classification Problems
Invited abstract in session TC-2: Infinite-dimensional optimization - Part II, stream Nonsmooth and nonconvex optimization.
Tuesday, 14:00-16:00Room: B100/7011
Authors (first author is the speaker)
| 1. | Lateef Jolaoso
|
| University of Southampton |
Abstract
The study of variational inequality problems (VIPs) continues to drive progress in the design of iterative methods, particularly under more relaxed conditions on the cost operator. While several methods have been developed for solving VIPs under monotone and pseudomonotone conditions, recent research has shifted focus toward tackling the more general case of quasimonotone operators. In this paper, we propose a novel double inertial acceleration algorithm that integrates self-adaptive and relaxation strategies to solve quasimonotone VIPs. We establish both weak convergence and linear convergence rate results under mild assumptions. To validate the efficiency of the proposed method, we conduct extensive numerical experiments, which demonstrate its superior performance compared to several existing methods in the literature. Additionally, we explore an application in machine learning by reformulating the support vector machine (SVM) classification problem as a VIP, thereby highlighting the practical relevance and applicability of our approach.
Keywords
- Non-smooth optimization
- Applications of continuous optimization
- Monotone inclusion problems
Status: accepted
Back to the list of papers