Abstract: Tis paper presents two-layer reinforcement learning method for multi-agent cooperation. This method is mainly realized by adding two-layer reinforcement learning unit to every agent. A experiment that three agents raise a round object by cooperation has been done. The result shows that the agents cooperate better than the agents using traditional reinforcement learning .

摘要： 提出了多智能体协作的两层强化学习方法。该方法主要通过在单个智能体中构筑两层强化学习单元来实现,将该方法应用于3个智能体协作抬起圆形物体的计算机模拟中,结果表明比采用传统强化学习方法的智能体协作得更好。

关键词: 强化学习, Q学习, 多智能体, 协作

CLC Number: 

• TP242