RoboCup仿真中多Agents的最优局部配合

为了促进多智能体技术发展，本提出了一种基于协调图和变量排除算法的局部合作策略体系，协调图用来表现一个系统的协调需求，变量排除算法用来求出最优配合，给出了理论上的证明和实际的应用．在RoboCup仿真中。该策略体系从微观上对agents的合作行为进行指导，在局部形成最优的合作策略，有效提高了传球的成功率和球队的攻击力。在实际应用中也取得了较好的效果．

An Optimal Local Collaboration of Multi-Agent in RoboCup Simulation

In order to promote the technology of Multi-Agent,this paper introduces a local strategy of collaboration based on coordination graph and variable elimination algorithm.coordination graph represents the coordination requirements of a system,variable elimination algorithm compute the optimal local collaboration.In RoboCup simulation,this architecture instructs the collaborate action in the detail,makes the best decisions locally,and improve the success rate of passing and the aggression of the whole team efficiently.The practical effect is very good.