摘要
再励学习(Reinforcement Learning,RL)是一种成功地结合动态编程和控制问题的机器智能方法,它将动态编程和有监督学习方法结合到机器学习系统中,通常用于解决预测和控制两类问题。提出了以矢量形式表示的评估函数,为了实现多维再励学习,用一专门的神经网络(Q网络)实现评判网络,研究其在移动机器人行为规划中的应用。
Reinforcement Learning(RL) is an approach to machine intelligence that combines two problems of Dynamic Programming and Control successfully.It combines the fields of dynamic programming and supervised learning to yield powerful machine-learning systems.RL has traditionally been used to solve problems of prediction and control.This paper proposes an evaluation function expressed in a vector form in order to realize multi-dimensional reinforcement learning.Q-learning,A special neural network (Q-net) is proposed to realize critic networks.at the end,we investigate the application of a Reinforcement learning in behavior planning.
出处
《工业控制计算机》
2009年第8期58-59,共2页
Industrial Control Computer
基金
海南省教育厅自然科学基金资助项目(Hj2009-134)
关键词
再励学习
神经网络
智能机器人
行为规划
应用
reinforcement learning,neural networks,intelligent robot,behavior planning,application