基于MAS的协作Agent间的联盟策略研究

Agent是一个在分布式系统或者协作式系统中能持续自主发挥作用的实体,它具有自主性、主动性、反应性和智能性的特点.随着单个Agent研究的逐渐深入,多Agent系统越来越受到关注,多agent机器人系统(MAS)是目前研究的热点.个体Agent之间互相合作、协作和相互服务,可以共同构成一个功能强大的多Agent系统(多智能体系统-MAS).MAS可以由多个智能和能力较低或较单一的Agent组成,也可以以几个较复杂的Agent成员为基础,结合其他简单Agent成员共同组成.本文首先介绍了Agent的概念、特性和结构模型,通过分析Shapley和Postman算法的优缺点对多Agent在协作过程中的联盟策略进行了研究.

Study on the Coalition Strategy of the Coordinate Agent Based on the MAS

Agent,continuously playing the active function in the distributed artificial intelligence and coordinative system,is featured with autonomy,activity,reactivity and intelligence.With the Single-agent research developing into a higher level,multi-agent system arouses more and more attention,and MAS has become the current focus in the research field.The mutual coordination and mutual service significantly contribute to the forming of greatly functional multi-agent system.MAS could be made up of either several agents with relatively low and unitary intellectual ability or a couple of complex agents combining with some other simplistic agents.Beginning with the conception, attributes and structural model of agent,this paper systematically studies the coordinative strategies during the process of agents cooperating through analyzing the Shapley and Postman algorithm on their advantages and disadvantages respectively.