贸易经济系学术研讨会系列2023年第1讲

来源:贸易经济系 发布时间:2023-05-11

主 题:Distributed Algorithm Design for Noncooperative Games

主讲人:邓镇华(中南大学副教授)

时 间:2023年5月11日(周四)上午10:00

地 点:明商1008会议室

协调人:陈志远

语 言:中/英文


讲座摘要:

Noncooperative games (NGs) widely emerge in numerous fields, such as economic markets, social networks, sensor networks, and smart grids. The genralized Nash equilibrium (GNE) is an interesting strategy profile for NGs, whose seeking has attracted increasing interest. In this talk, firstly, we present a distributed projection-based seeking strategy for nonsmooth NGs, where the payoff functions are nondifferentiable. It is proved that the strategy is initial-decision free, and can seek the exact variational GNE of the games. Then we investigate the NGs of second-order players, since, with the intellectualization of cyber-physical systems, physical systems have been used to perform distributed game tasks (e.g., the electricity market games of generators, the attack-defense games of UAVs). In the problem, the players have second-order dynamics, and they are subject to local and coupling inequality constraints. For the problem, we design a distributed primal-dual-based strategy, with the help of gradient descent and state feedback. By rigorous analysis, we prove that the strategy is globally asymptotically convergent. Our two methods are also verified via electricity market games.


主讲人简介:

邓镇华,中南大学自动化学院副教授、博士生导师。本、硕、博分别就读于大连海事大学、中南大学和中科院系统所。主要从事分布式优化与博弈、多智能体控制、智能电网、集群自主无人系统等方面的研究工作。近年来在Automatica、IEEE Trans.等期刊发表SCI论文30余篇;主持国家级项目2项、省部级项目1项。

人大商学院新闻网版权与免责声明:

① 凡本网未注明其他出处的作品,版权均属于人大商学院,未经本网授权不得转载、摘编或利用其它方式使用上述作品。已经本网授权使用作品的,应在授权范围内使用,并注明“来源:人大商学院”。违反上述声明者,本网将追究其相关责任。

② 凡本网注明其他来源的作品,均转载自其它媒体,转载目的在于传递更多信息,并不代表本网对其负责。

③ 有关作品内容、版权和其它问题请与本网联系。

※ 联系方式:中国人民大学商学院宣传信息事务办公室 邮箱:media@rmbs.ruc.edu.cn

官方微信 中国人民大学商学院 86-10-82509171 rmbs@rmbs.ruc.edu.cn

©中国人民大学商学院 版权所有 京ICP备05066828号-1