基于强化学习的模糊自适应控制器
来源期刊：中南大学学报(自然科学版)1998年第2期
论文作者：马莉 蔡自兴
文章页码：70 - 73
关键词：强化学习;模糊控制;神经网络
Key words：reinforcement learning; fuzzy control; neural networks
摘 要：提出了一个强化学习系统中模糊自适应控制器网络结构及其有关算法的改进.并在此基础上给出了二阶欠阻尼系统和强非线性系统的强化学习控制仿真结果.仿真结果表明,基于强化学习的模糊自适应控制器可以对一类复杂系统实现自学习控制,达到令人满意的控制精度.最后,作者还对进一步研究的问题进行了探讨.
Abstract: Reinforcement learning is a self-learning control approach which can adjust control rules on line to reach the desired outputs through interaction between the inference engine and environment.This paper provides an architecture of a fuzzy adaptive controller network in the RL system and some related-algorithm modification.Some computer simulation shows that the controller proposed can realize the selflearning with satisfied accuracy for a class of complicated system.
