作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2008, Vol. 34 ›› Issue (11): 219-221. doi: 10.3969/j.issn.1000-3428.2008.11.079

• 人工智能及识别技术 • 上一篇    下一篇

共享经验的多主体强化学习研究

焦殿科1,石 川2   

  1. (1. 辽宁工业大学计算机科学与工程学院,锦州 121001;2. 北京邮电大学北京市智能软件与多媒体重点实验室,北京 100088)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-06-05 发布日期:2008-06-05

Research on Multi-agent Reinforcement Learning with Sharing Experience

JIAO Dian-ke1, SHI Chuan2   

  1. (1. College of Computer Science & Engineering, Liaoning University of Technology, Jinzhou 121001; 2. Beijing Key Laboratory of Knowledgeware and Multimedia, Beijing University of Posts and Telecommunications, Beijing 100088)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-06-05 Published:2008-06-05

摘要: 合作多主体强化学习的关键问题在于如何提高强化学习的学习效率。在追捕问题的基础上,该文提出一种共享经验的多主体强化学习方法。通过建立合适的状态空间使猎人共享学习经验,根据追捕问题的对称性压缩状态空间。实验结果表明,共享状态空间能够加快多主体强化学习的过程,状态空间越小,Q学习算法收敛越快。

关键词: 合作多主体, 强化学习, Q学习算法, 状态空间

Abstract: How to improve the efficiency of reinforcement learning is the key problem of reinforcement learning with multi-agent collaboration. This paper proposes a method of multi-agent reinforcement learning with sharing experience based on the research to pursuit problem. By applying this method the hunters can share the learning experience through constructing the appropriate state space. It further compresses the state space according to the symmetry character of pursuit problem. Experimental results show that sharing state space can expedite the process of multi-agent reinforcement learning. The smaller the state space is, the faster Q learning algorithm convergence will be.

Key words: multi-agent collaboration, reinforcement learning, Q learning algorithm, state space

中图分类号: