期刊文献+
任意字段
题名或关键词
题名
关键词
文摘
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息

一种基于增强学习的自适应控制方法 被引量：4

A novel adaptive control algorithm based on reinforcement learning
下载PDF
导出
摘要 针对模型未知时变非线性对象的控制问题 ,提出一种直接的自适应控制策略。该策略基于径向基神经网络并结合增强学习的自调节能力 ,无需知道控制对象的动态特性 ,而是通过在线试错在控制过程中不断积累与问题相关的信息 。 Adirectadaptivecontrolschema for model unknown time dependent dynamical plant is presented.Combiningtheabilityofradialbasedfunction network and the self tune property of reinforcement learning, it needs not to know the priori knowledge of the plant′s dynamical property. The knowledge related to the problem is accumulated gradually by trial and error method and then the acceptable optimal solution can be obtained.
作者 顾冬雷 陈卫东 席裕庚
机构地区 上海交通大学自动化研究所
出处 《控制与决策》 EI CSCD 北大核心 2002年第4期473-475,479,共4页 Control and Decision
基金 国家自然科学基金项目 (6 0 10 5 0 0 5 )
关键词 增强学习 径向基神经网络 自适应控制 reinforcement learning radial based function network adaptive control
分类号 TP24 [自动化与计算机技术—检测技术与自动化装置]
  • 相关文献

参考文献5

  • 1[1]D A Linkens, H O Nyongesa. Learning systems inintelligent control: An appraisal of fuzzy, neural and genetic algorithm control applications[J]. IEE Proc on Control Theory and Appl,1996,143(4):367-386.
  • 2[2]Kaelbling L P, Littman M L, Moore A W. Reinforcement learning: A survey[J]. J of Artificial Intell Res,1996,4:237-277.
  • 3[3]Barto A G, Sutton R S, Anderson C W. Neuronlike adaptive elements that can solve difficult learning control problems[J]. IEEE Trans on Syst, Man &Cybern,1983,13(5):834-846.
  • 4[4]Kimura H, Kobayashi S. An analysis of actor/critic algorithms using eligibility traces: Reinforcement lear-ning with imperfect value function[A]. 15th Int Conf on Machine Learning[C]. Madison,1998.278-286.
  • 5[5]Clouse J A, Utogoff P E. A teaching method for reinforcement learning[A]. Proc of the 9th Int Conf on Machine Learning[C]. San Mateo,1992.93-101.

同被引文献29

  • 1文锋,陈宗海,卓睿,周光明.连续状态自适应离散化基于K-均值聚类的强化学习方法[J].控制与决策,2006,21(2):143-147. 被引量：7
  • 2陈宗海,文锋,聂建斌,吴晓曙.基于节点生长k-均值聚类算法的强化学习方法[J].计算机研究与发展,2006,43(4):661-666. 被引量：13
  • 3[1]Sutton R S, Barto A G. Reinforcement Learning: An Introduction [M]. MIT Press, 1998.
  • 4[2]Barto A G, Sutton R S, Anderson C W. Neuronlike adaptive elements that can solve difficult learning control problems [J]. IEEE Transactions on SMC,1983, 13(5) :834 ～846.
  • 5[3]Xu X, He H-G, Hu D W. Efficient reinforcement learning using recursive least-squares methods [ J ]. Journal of Artificial Intelligence Research, 2002,16:259 ～292.
  • 6[4]Kimura H, Kobayashi S. An analysis of actor/critic algorithms using eligibility traces: reinforcement learning with imperfect value functions [A]. 15th Int. Conf. on Machine Learning [C]. Madison, 1998. 278～286.
  • 7A.G.Barto,S.J.Bradtke,S.P.Sign.Learning to act using real-time dynamic programming[Master dissertation].Amherst:Department of Computer Science,University of Massachusetts,1995
  • 8C.S.Lin,H.Kim.Selection of learning parameters for CMAC-based adpative critic learning.IEEE Trans.Neural Networks,1999,6(3):642～647
  • 9A.J.Smith.Application of the self-organising map to reinforcement learning.Neural Networks,2002,15(8/9):1107～1124
  • 10I.Menache,S.Mannor,N.Shimkin.Basis function adaptation in temporal difference reinforcement learning.http://www.ee.technion.ac.il/people/shimkin/PREPRINTS/BasisAdaptation-Dec03.pdf,2003

引证文献4

二级引证文献17

控制与决策
2002年 第4期

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部