不完全信息下的非通信多Agent协作控制研究

Study on Non-communicative Multi-agent Collaboration under Incomplete Information

针对分布式多Agent系统(MAS)的动态不确定性、信息不完全性以及非信讯的多智能体协作特点,引入了Agent递减算法,并据agent行为预测的具体情况对其加以调整,而后结合提出的时空表控制方式,解决了不完全信息下的多Agent的非通信协作问题,并提高了协作效率。最后,以足球机器人RoboCup仿真比赛系统实际应用为例,验证了该方法的可行性和有效性。

Aiming at the robotic soccer simulation system which holds the characteristic of dynamic, uncertainty and non-communicative multi-agent collaboration, first, agent descending arithmetic is imported and adjusted it according to the prediction of the agent action in multi-agent system, then combind with the way of space-time table controller to solve the non-communicative problems of multi-agent and improved the efficiency of cooperation. Finally, the feasibility of this algorithm is verified by robotic soccer simulation system tests.