关键词: 多智能体系统, 局部合作, Q-学习, 过程分层

Abstract: Many multi-agent Q-learning problems can not be solved because of the number of joint actions is exponential in the number of agents.Based on the study of the cooperation in MAS in reinforcement learning and on the basis of the research in the system logic，this paper puts forward the hierarchical regional cooperation reinforcement learning based on learning process.By studying the knowledge of Agent reinforcement learning and improving the multi-Agent study efficiency，the performance of the regional cooperation reinforcement learning is further enhanced，combining with the mission action based on joint action and potential field model so as to solve the dimensional disaster in state space of reinforcement learning.This algorithm is used in a subtask of robot soccer and its effectiveness is validated by experiments.

Key words: Multi-Agent Systems（MAS）, regional cooperative, Q-learning, process stratification