摘要
强化学习是一种重要的机器学习方法,然而在实际应用中,收敛速度缓慢是其主要不足之一。为了提高强化学习的效率,提出了一种并行强化学习算法。多个同时学习,在各自学习一定周期后,利用D-S证据利用对学习结果进行融合,然后在融合结果的基础上,各进行下一周期的学习,从而实现提高整个系统学习效率的目的。实验结果表明了该方法的可行性和有效性。
Reinforcement learning is an important machine learning method.However,slow convergence has been one of main problem in practice.To improve the efficiency of reinforcement learning,this paper proposes parallel reinforcement learning algorithm.There are multiple agents in learning system.In a learning episode ,each agent learns independently.After a learning episode, the results of all agents are fused based on D-S evidence theory so as to achieve common result, which are shared by all agents in next learning episode.Experiments show the feasibility and efficiency of the algorithm.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第34期25-28,52,共5页
Computer Engineering and Applications
基金
国家"十一五"科技支撑计划重大项目资助No.2006BAD03A02~~
关键词
并行算法
强化学习
Q-学习
D—S证据理论
路径规划
parallel algorithms
reinforcement learning
Q-learning
D-S evidence theory
path plan