基于混沌BP算法的TS型模糊神经网络的应用研究.pdfVIP

th Proceedings of the 25 Chinese Control Conference 7-11 August, 2006, Harbin, Heilongjiang 基于混沌BP算法的T-S型模糊神经网络的应用研究 王科俊，唐 墨，张 燕 哈尔滨工程大学 自动化学院, 哈尔滨 150001 E-mail: heukejun@, {tmtangmo, zylavender}@163.com 摘 要: T-S型模糊神经网络对非线性函数具有良好的逼近能力，其学习算法多采用BP算法。本文将一种新型的 非线性混沌映射引入到T-S型模糊神经网络的参数学习算法中。将混沌优化算法与BP算法相结合，使模糊神经 网络克服了容易陷入局部极小值的缺点。并对非线性和混沌系统进行建模，仿真结果表明，用混沌BP学习算法 进行参数更新，能够减小误差，提高网络性能。 关键词: 模糊神经网络, BP算法, 混沌 Application and Research of T-S Fuzzy Neural Network Based on Chaotic BP Method Kejun Wang, Mo Tang, Yan Zhang Harbin Institute of Technology, Harbin 150001, China E-mail: heukejun@, {tmtangmo, zylavender}@163.com Abstract: In this paper a new chaotic BP algorithm is studied. That is introducing a new type of nonlinear chaotic map into the learning algorithm of T-S fuzzy neural network. The combination of chaos and BP learning method make fuzzy neural network escape from getting in local minimization value. Then approximate nonlinear and chaotic systems using the new learning algorithm. Simulation result show that the transform could improve the property of neural network, which is capable of reduce error and have better generalization performance. Key Words: Fuzzy Neural Networks, BP Algorithm, Chaos 射具有遍历性、随机性等特点。混沌BP算法就是将 1 引言(Introduction) 混沌映射的遍历性与梯度法相结合，通过引入混沌 来跳出局部极小。目前，已有的混沌优化方法[1]-[6]多 模糊神经网络融合了神经网络与模糊系统二者 采用Logistic映射作为混沌序列发生器。文[1]中，通 的优势，成为模糊系统及其控制的一个重要分支。 过Logistic 映射函数来