一种基于量子遗传算法的扩展ts模型辨识-控制与决策.pdf

第28 卷 第8 期 控 制 与 决 策 2013 年 8 月 Vol. 28 No. 8 Control and Decision Aug. 2013 文章编号: 1001-0920 (2013) 08-1268-05 一种基于量子遗传算法的扩展T-S 模型辨识 李 浩, 李士勇 (哈尔滨工业大学控制科学与工程系，哈尔滨150001) 摘 要: 在传统T-S 模型的基础上, 提出一种扩展T-S 模型. 该模型由一组模糊规则组成, 由规则前件实现输入空间 的划分, 将成员函数及其函数变换引入规则后件以实现对输入子空间的非线性映射. 对于该模型的建立, 使用改进量 子遗传算法优化规则前件, 递推最小二乘法确定规则后件参数. 通过对两个典型非线性系统辨识, 仿真结果表明了该 模型可以显著提高辨识精度, 且具有很好的泛化性能. 关键词: 扩展T-S 模型；模糊规则；成员函数；量子遗传算法；递推最小二乘法 中图分类号: TP273 文献标志码: A An expanded T-S model identification based on quantum genetic algorithm LI Hao, LI Shi-yong (Department of Control Science and Engineering，Harbin Institute of Technology ，Harbin 150001 ，China. Correspondent ：LI Hao ，E-mail ：hblh198095@163.com) Abstract: An expanded T-S model is proposed based on the conventional T-S model. This model is comprised of a set of fuzzy rules. According to the premise part of the rules, the input space can be partitioned, and the membership values and their transformations are introduced in the consequent part of the rules to express the nonlinear mapping relation in the input subspace. To construct the model, the improved quantum genetic algorithm is used to optimize the premise part of the rules, and the recursive least squares method is used to determine the parameters in the consequent part of the rules. Through the identification of two nonlinear systems, simulation results show that the proposed model can improve the approximation accuracy and have excellent generalization ability. Key words: expanded T-S model ；fuzzy rule ；membership value ；quantum genetic algorithm ；recursive least squares method 0 引引引 言言言 的建立, 该文献