arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1306.2918cs.GTcs.LGmath.PR
收藏

具有动作集限制的强化学习

Reinforcement learning with restrictions on the action set

Aix-Marseille University(艾克斯-马赛大学);CNRS(法国国家科学研究中心);EHESS(法国高等社会科学研究院)

Mario Bravo, Mathieu Faure

摘要:

考虑一个2人标准型博弈,随着时间的推移不断重复。我们引入了一个自适应学习过程,在每个阶段,玩家只观察自己实现的收益。我们假设代理人不知道自己的支付函数,也没有其他球员的信息。此外,我们假设他们有自己的行动集的限制,这样,在每个阶段,他们的选择是有限的,他们的行动集的一个子集。我们证明了经验分布的发挥收敛到纳什均衡的零和和潜在的游戏,游戏中一个球员有两个行动。

英文摘要:

Consider a 2-player normal-form game repeated over time. We introduce an adaptive learning procedure, where the players only observe their own realized payoff at each stage. We assume that agents do not know their own payoff function, and have no information on the other player. Furthermore, we assume that they have restrictions on their own action set such that, at each stage, their choice is limited to a subset of their action set. We prove that the empirical distributions of play converge to the set of Nash equilibria for zero-sum and potential games, and games where one player has two actions.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑