基于模糊Q-学习的多智能体协作策略研究  

Cooperation strategy among multi-agent based on fuzzy Q-learning

作　　者：郝宗波[1] 洪炳镕[1] 周彤[1] 

机构地区：[1]哈尔滨工业大学计算机科学与技术学院,黑龙江哈尔滨150001

出　　处：《哈尔滨工业大学学报》2004年第7期931-933,共3页Journal of Harbin Institute of Technology

基　　金：国家高技术研究发展计划资助项目(863-2001AA422270).

摘　　要：研究了对机器人足球中的非控球队员同控球队员的协作问题.机器人根据当前状态下机器人和球的位置信息决定目标位置.对目标位置的确定进行了Q学习,对于场地的划分和角色的分配,采用了模糊理论,并对学习结果进行了仿真试验.The cooperation of the non-control robot with the control robot is studied. The destination of the robot is determined by the situation of the robots and ball, which is learned with the Q-learning. The partition of the region and the assignment of the roles are preformed by fuzzy theory. And the simulation experiment is made to verify the result of the learning.

关 键 词：协作 Q学习 模糊 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]