种自适应模糊Actor—Critic学习.pdfVIP

See discussions, stats, and author profiles for this publication at: /publication/265326218 A kind of adaptive fuzzy Actor-Critic learning Article in Kongzhi yu Juece/Control and Decision · September 2006 CITATIONS 2 READS 38 3 authors: Xuesong Wang China University of Mining Technology 107 PUBLICATIONS 300 CITATIONS SEE PROFILE Yuhu Cheng China University of Mining Technology 52 PUBLICATIONS 182 CITATIONS SEE PROFILE Jianqiang Yi Chinese Academy of Sciences 306 PUBLICATIONS 2,456 CITATIONS SEE PROFILE All content following this page was uploaded by Jianqiang Yi on 21 May 2015. The user has requested enhancement of the downloaded file. 第 21卷 第9期 Vo1．21 No．9 控 制 与 Control and 2006年 9月 Sep．2006 文章编号 ：1001—0920(2006)09—1068—05 一 种 自适应模糊Actor—Critic学习 王雪松 ，程玉虎 ，易建强 (1．中国矿业大学 信息与电气工程学院，江苏 徐州 221008；2．中国科学院 自动化研究所，北京 100080) 摘 要 ：提 出一种基于模糊 RBF 网络的 自适应模糊 Actor—Critic学习．采用一个模糊 RBF神经 网络 同时逼近 Actor 的动作函数和Critic的值函数，解决状态空间泛化中易出现的“维数灾”问题．模糊RBF网络能够根据环境状态和被 控对象特性的变化进行网络结构和参数的自适应学习，使得网络结构更加紧凑，整个模糊Actor—Critic学习具有泛化 性能好、控制结构简单和学习效率高的特点．Mountain Car的仿真结果验证 了所提方法的有效性． 关键词：Actor—Critic学习；模糊推理系统；RBF网络；泛化 中图分类号：TP18 文献标识码：A A Kind of Adaptive Fuzzy Actor-Critic’Learning W ANG Xue—song ，CH ENG Yu—hu ，y Jian—qiang (1．School of Information and Electrical Engineering，China University of Mining and Technology，Xuzhou 221008， China：2．Institute of Automation，Chinese Academy of Sciences，Beijing 100080，China．Correspondent：WANG Xue—song，E—mail：wangxuesongcumt@163．com) Abstract：An adaptive fuzzy Actor·Critic reinforcement learning based on a fuzzy radial basis function network is proposed，which can solve the ‘curse of dimensionality’problem caused by state space generalization．A fuzzy RBF network is used to approximate both the action function of Actor and the value function of Critic simultaneously． The fuzzy RBF network is able to adjust its structure and parameters in an adaptive way with a self—organizing approach according to the change of environment state