基于换位思考模型的多智能体协作研究及在RoboCup中的应用  

Research in Multi Agent System Based on Change Role and Application in RoboCup

作　　者：王浩[1] 尚丽[1] 方宝富[1] 李龙[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230001

出　　处：《小型微型计算机系统》2009年第5期959-962,共4页Journal of Chinese Computer Systems

基　　金：上海市机械自动化及机器人重点实验室开发课题(Z0802)资助

摘　　要：机器人世界杯(RoboCup)是一个典型的多智能体系统.为了提高多智能体协作的效率,提出一种新的基于换位思考模型的多智能体协作研究方法.首先,教练智能体获取仿真比赛环境中球员智能体的无噪音信息,对所有队友智能体建模;然后,应用高斯分布计算队友智能体的当前行为模式,并把当前模式反馈给仿真环境;最后,球员智能体根据换位思考模型计算得到的模式做出相应决策.该模型已经应用于HfutEngine2D仿真球队中,在RoboCup仿真比赛中获得2007年中国公开赛亚军,2008年机器人世界杯第7的好成绩.In order to enhance succeed of cooperation in Multi Agent System, this paper proposes a new approach to do research on Multi Agent System based on＂ Change Role＂. In this scene, an autonomous coach agent is able to get the current information of all teammates without noise, which can be modeled to compose patterns of teammates. Firstly coach gathers data from noisy environment to identify pattern of player agent. Then compute the probability of pattern compared with current situation by statistical calculations. Finally make successful decision according to the result of＂ change role＂ with correlative teammates. The new approach based on ＂Change Role＂ is tested In ChinaOpen2007 in Robot Soccer Game, Simulation 2D team-HfutEngine2D, and get 2nd. In the past World RoboCup2008, we got 7th.

关 键 词：换位思考模型 多智能体协作 智能体建模 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]