| 模糊神经网络的混沌优化算法设计 |
| |
| 引用本文： | 邹 恩,李祥飞,张泰山. 模糊神经网络的混沌优化算法设计[J]. 控制理论与应用, 2005, 22(4): 578-582 |
| |
| 作者姓名： | 邹 恩 李祥飞 张泰山 |
| |
| 作者单位： | 株洲工学院,电气工程系,湖南,株洲,412008;中南大学,信息科学与工程学院,湖南,长沙,410083;株洲工学院,电气工程系,湖南,株洲,412008;中南大学,信息科学与工程学院,湖南,长沙,410083 |
| |
| 基金项目： | supportedbyHunanProvinceNaturalScienceFoundation(01JJY3029)，ChinaNationalPackagingCorporationScienceandTechnologyFoundation(05ZBKJA003). |
| |
| 摘 要： | 提出了一种基于混沌变量的多层模糊神经网络优化算法设计.离线优化部分采用混沌算法,将混沌变量引入到模糊神经网络结构和参数的优化搜索中,使整个网络处于动态混沌状态,根据性能指标在动态模糊神经网络中寻找较优的网络结构和参数.在线优化部分采用梯度下降法,把混沌搜索后得到的参数全局次优值作为梯度下降搜索的初始值,进一步调整模糊神经网络的参数,实现混沌粗搜索和梯度下降细搜索相结合的优化目的,能较快地找到全局最优解.最后对二阶延迟系统进行仿真,结果表明混沌优化方法控制精度高、超调小、响应快和鲁棒性强.
|
| 关 键 词： | 模糊神经网络 优化 混沌变量 梯度下降法 |
| 文章编号： | 1000-8152（2005）04-0578-05 |
| 收稿时间： | 2002-12-03 |
| 修稿时间： | 2002-12-032004-12-08 |
Chaos optimization algorithm design for fuzzy neural network |
| |
| ZOUEn,LI Xiang-fei,ZHANG Tai-shan. Chaos optimization algorithm design for fuzzy neural network[J]. Control Theory & Applications, 2005, 22(4): 578-582 |
| |
| Authors: | ZOUEn LI Xiang-fei ZHANG Tai-shan |
| |
| Affiliation: | Department of Electrical Engineering ,Zhuzhou Institute of Technology,Zhuzhou Hunan412008,China; Information Science & Engineering Institute ,Central South University,Changsha Hunan410083,China |
| |
| Abstract: | An optimization algorithmdesign based on chaotic variable is proposed for multilayer fuzzy neural network.Off-line optimization uses chaos algorithmand chaos variables are applied to searchfor networkstructure and parameters ,in which thenetworkis in dynamic chaos state .An approximate optimal network structure and parameters are found from dynamic networkaccordingto performance index.On-line optimization uses gradient descent algorithmand the initial values of gradient descentsearching are parameters approximately global optimal values fromchaos searching,the parameters of fuzzy neural network are fur-ther adjusted.The global optimal values of networkare searched quickly by means of combination of chaos global searchingand gra-dient descent local searching.Finally,second order delaysystemis simulated,andthe results showthat the chaos optimal control is ofhigh precision,small overshoot ,fast response and good robustness . |
| |
| Keywords: | fuzzy neural network optimization chaotic variables gradient descent algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《控制理论与应用》浏览原始摘要信息 |
|
点击此处可从《控制理论与应用》下载全文 |
|
