报告人：雷金龙(同济大学)

报告时间：2019年11月4日（周一）9:00-10:00

报告地点：南楼N224

报告摘要: The distributed computation of equilibria has seen growing interest in a broad collection of networked problems. This talk focus on two classes of stochastic Nash games: a potential stochastic Nash game where each player solves a parameterized stochastic convex program, and a misspecified generalization where the player-specific stochastic program is complicated by a parametric misspecification. Two asynchronous inexact proximal BR schemes are designed, where in each iteration, a single player is randomly chosen to compute an inexact proximal BR solution with delayed rival information while the other players keep their strategies invariant. In the misspecified regime, each player possesses an extra estimate of the misspecified parameter. By imposing suitable conditions on the inexactness sequences, the iterates produced by both schemes are shown to converge almost surely to a connected subset of the set of Nash equilibria. Besides, the associated gap function is shown to converge to zero in mean. These statements can be extended to allow for accommodating weighted potential games and generalized potential games.

个人简介： 雷金龙，同济大学电子与信息工程学院“青年百人B计划”特聘研究员。主要研究方向是不确定信息下多智能体网络优化与非合作博弈的分析与设计理论，包括随机纳什博弈、随机非凸优化、随机逼近、分布式估计与分布式优化等。2011年于中国科学技术大学获得学士学位，2016年于中国科学院数学与系统科学研究院获得理学博士学位.2016年8月- 2019年9月，在美国宾夕法尼亚州立大学工业工程系从事博士后研究。在运筹学与控制域 Operations Research, IEEE Transcations on Automatic Control, Mathematics of Operations Reseach, SIAM Journal on Control and Optimization 等期刊上发表多篇论文。