Multi-type Multi-agent Non-cooperative Systems in the High Population Regime

在多智能体动态系统中,分析各智能体在不对称信息下的非合作随机博弈是十分困难的。因为每个智能体信念的产生通常涉及无限递归,这增加了寻找均衡的难度。面对这个挑战,我们发现增加智能体的数目可以弱化个体交互对系统的影响,这就是所谓的平均场博弈(Mean Field Games, MFGs)。
第三期 IEEE TNSE 杰出讲座系列活动,我们邀请到美国伊利诺伊大学厄巴纳-香槟分校 Tamer Başar 教授分享使用 MFGs 方法,在多智能体动态系统中进行决策的基本原理、分析方法与未来研究方向。
通过活动行(http://hdxu.cn/66t0I)或哔哩哔哩参加(http://live.bilibili.com/21845454)。
IEEE TNSE 杰出讲座系列由 IEEE TNSE 期刊和深圳市人工智能与机器人研究院(AIRS)联合主办,香港中文大学(深圳)、网络通信与经济学实验室(NCEL)、IEEE 联合支持。该系列活动旨在汇聚网络科学与工程领域的国际顶级专家学者分享前沿科技成果。
-
黄建伟香港中文大学(深圳)校长讲座教授、理工学院副院长、AIRS 副院长兼群体智能中心主任、IEEE TNSE主编、IEEE Fellow、AAIA Fellow执行主席
-
Tamer Başar美国伊利诺伊大学厄巴纳-香槟分校Swanlund名誉主席、电气与计算机工程系CAS名誉教授、IEEE Fellow、IFAC Fellow、SIAM FellowMulti-type Multi-agent Non-cooperative Systems in the High Population Regime
Tamer Başar 于1981年加入伊利诺伊大学厄巴纳-香槟分校(UIUC),目前是 Swanlund 名誉主席、电气与计算机工程系 CAS 名誉教授、CSL 和 ITI 研究教授。他还是 Illinois@Singapore 的执行理事。在 UIUC,他曾担任 CAS 主任(2014-2020)、工程学院临时院长(2018)和 Beckman 研究所临时主任(2008-2010)。他是美国国家工程院院士,IEEE、IFAC、SIAM Fellow;并曾担任 IEEE CSS、ISDG、AACC 的主席。多年来,他获得了多个奖项和荣誉,包括 IEEE CSS(Bode Lecture Prize)、IFAC(Quazza Medal)、AACC(Bellman Award)和ISDG(Isaacs Award)的最高奖项,IEEE 控制系统技术领域奖,获母校耶鲁大学 Wilbur Cross 奖章,以及多个国际名誉博士学位和教授职位。他于2004-2014年期间担任 IFAC Journal Automatica 的主编,目前是多个丛书的主编。他在系统、控制、通信、优化、网络和动态博弈等领域做出了巨大贡献,他现在的研究兴趣包括随机团队、博弈和网络;平均场博弈;多智能体系统和学习;多智能体系统中的激励;数据驱动的分布式优化;流行病建模和网络控制;能源系统;和网络物理系统等。
Perhaps the most challenging aspect of research on multi-agent dynamical systems, formulated as non-cooperative stochastic games with asymmetric dynamic information, is the presence of strategic interactions among agents, with each one developing beliefs on others in the absence of shared information. This belief generation process involves what is known as second-guessing phenomenon, which generally entails infinite recursions, thus compounding the difficulty of obtaining an equilibrium. This difficulty is somewhat alleviated when there is a high population of agents, in which case strategic interactions at the level of each agent become much less pronounced. With some structural specifications, this leads to what is known as mean field games (MFGs), which have been the subject of intense research activity during the last fifteen years or so. Following a general overview of fundamentals of MFGs approach to decision making in multi-agent dynamical systems, the talk will introduce a framework where the agents are partitioned into finitely-many populations with an underlying graph topology, with each population having a high number of indistinguishable agents. Results on existence, uniqueness, and characterization of equilibria will be presented, along with learning such equilibria in model-free settings. The talk will conclude with a discussion of future research directions.