基于博弈论的多Agent协商模型

A game-theory-based Multi-agent negotiation model

在开放、动态、具有自适应性和自治性的多Agent系统(MAS)中,协商是多Agent系统实现协调、协作和解决冲突的关键．博弈论是使用严谨的数学模型研究冲突条件下最优决策问题的理论,本文以此为数学工具,基于对时间的限制,结合协商的历史知识,考虑各协商者在拥有不完全信息的情况下,提出了基于博弈论的多Agent协商模型(GTMANM),并给出了协商策略求解的算法描述,用以来解决分布式环境下不完全信息的多人协商决策问题,以获得最大的期望收益,为以后把博弈论中其他更复杂的博弈思想引入到多Agent协商中来奠定了基础．

In a open and dynamic Multi-agent system with adaptability and autonomy, negotiation is a key factor to resolve a problem of coordination, cooperation and conflict. Game theory lays the solid mathematical foundation for the research of coordination and collaboration. Make use of this precise math model, combine with the history knowledge and time restriction of the negotiation, as well as consider the incomplete information of negotiation, this paper provides A Game Theory Based Multi-Agent Negotiation Model(GT-MANM) , and gives the algorithm describation of the stritagy in order to resolve so many negotiation problems in Distribute Artificial Intelligence (DAI) with uncertain inform-ation to obtain the maximum expected utilities. which lays the solid foundation to bring the complexity thought of game theory into the Multi-Agent negotiation.