报告人：王洪助理教授

报告题目：An Operator Splitting Approach for Solving Multi-Stage Stochastic Generalized Nash Equilibrium Problem

报告摘要：In this talk, we consider a class of multi-stage stochastic generalized Nash equilibrium (SGNE) problem with expected-value cost functions, where the feasible decision sets of all players are coupled together by a globally shared affine constraint. Upon utilizing the variational SGNE (v-SGNE) as a refined solution, the problem is reformulated as that of identifying the zeros of a sum of monotone operators through the integration of nonanticipativity constraints, primal-dual analysis and an augmentation of variables. We first propose a v-SGNE searching algorithm based on forward-backward operator splitting methods for solving the two-stage SGNE. Then we demonstrate the convergence of the proposed algorithm for fixed step-sizes under some mild assumptions. Subsequently, we extend the proposed method and its convergence result to the multistage case. This is a joint work with Hailin Sun.

报告人简介：王洪，深圳技术大学大数据与互联网学院助理教授。本科和硕士毕业于002全讯白菜网，2017年博士毕业于香港理工大学。2017年至2020年任华为技术有限公司2012实验室高级工程师。2021年至今，任深圳技术大学大数据与互联网学院助理教授，从事分布式非光滑优化理论及其算法研究，研究成果发表在SIAM Journal on Optimization、Journal of Computational Mathematics和Numerical Algorithm等期刊，主持国家自然科学基金青年基金1项。

报告时间：2024年4月18号上午09:30-12:00

报告形式：腾讯会议；会议号：290-859-760

获取会议密码请联系：taohy163@163.com