基于人工神经网络的并行强化学习自适应路径规划  

Application of Parallel Reinforcement Learning Based on Artificial Neural Network to Adaptive Path Planning

作　　者：耿晓龙[1] 李长江[1] 

机构地区：[1]西北工业大学,西安710129

出　　处：《科学技术与工程》2011年第4期756-759,共4页Science Technology and Engineering

摘　　要：强化学习是通过对环境的反复试探建立起从环境状态到行为动作的映射。利用人工神经网络的反馈进行权值的调整,再与高学习效率的并行强化学习算法相结合,提出了基于人工神经网络的并行强化学习的应用方法,并通过实验仿真验证了迭代过程的收敛性和该方法的可行性,从而有效地完成了路径学习。Reinforcement learning is an important class of learning techniques that learns to perform a certain task through trial and error interactions with an knowledge-poor environment.By combining artificial neural network with parallel reinforcement learning,an applicable method of parallel reinforcement learning algorithm based on artificial neural network is proposed.Experimental results show that the method is effective.

关 键 词：并行强化学习 BP神经网络 路径规划 Q学习 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]