基于蒙特卡罗学习的多机器人自组织协作  

Self-organizing coordination of multi-robot based on Monte Carlo learning

作　　者：周彤[1] 洪炳镕[1] 朴松昊[1] 周洪玉[2] 

机构地区：[1]哈尔滨工业大学计算机科学与技术学院,哈尔滨150001 [2]哈尔滨理工大学机械动力学院,哈尔滨150080

出　　处：《计算机工程与应用》2007年第30期23-25,32,共4页Computer Engineering and Applications

基　　金：国家自然科学基金项目(the National Natural Science Foundation of China under Grant No.69985002);国家高技术研究发展计划(863)(the National High-Tech Research and Development Plan of China under Grant No.2006AA04Z259)。

摘　　要：强化学习是提高机器人完成任务效率的有效方法,目前比较流行的学习方法一般采用累积折扣回报方法,但平均值回报在某些方面更适于多机器人协作。累积折扣回报方法在机器人动作层次上可以提高性能,但在多机器人任务层次上却不会得到很好的协作效果,而采用平均回报值的方法,就可以改变这种状态。本文把基于平均值回报的蒙特卡罗学习应用于多机器人合作中,得到很好的学习效果,实际机器人实验结果表明,采用平均值回报的方法优于累积折扣回报方法。Reinforcement learning is an effective way for accomplishing task in multi-robot system.While much of the work has focused On optimizing discounted cumtilative reward,optimizing average reward is sometimes a more suitable criterion for multi-robot coordination.Learning algorithms based on discounted rewards,such as Q learning,can attain a well result at the action-level,but it cannot perform well at the task-level.However,learning methods based on average reward,such as the Monte Carlo algorithm,are capable of achieving the optimal result through cooperation at the task-level.Real robot experiment shows that the algorithm adopting the average reward is superior to the one adopting the discounted cumulative reward.

关 键 词：强化学习 多机器人 协作 蒙特卡罗学习 Q学习 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]