基于遗传算法的非线性系统辨识的研究  

Research of nonlinear system identification based on genetic algorithm

作　　者：刘正龙[1] 杨艳梅[2] 罗玉军[1] 

机构地区：[1]川北医学院基础医学院,南充637007 [2]西华师范大学数学与信息学院,南充637002

出　　处：《黑龙江大学自然科学学报》2014年第3期416-420,共5页Journal of Natural Science of Heilongjiang University

基　　金：四川省教育厅自然科学基金资助项目(12ZB040);四川省教育厅教育发展研究中心基金资助项目(CJF10019)

摘　　要：针对线性、时不变和具有不确定参数的对象进行辨别和控制研究,其辨别器和控制器的确保辨识系统全局稳定的自适应函数调参规律和结构选优组成等,都是构建在线性系统理论之上。由于非线性系统的辨别和自适应控制一直难于找到相应的数学方法,因此提出一种非线性系统数学模型构成系统模型辨识的新方法,将组合优化问题由非线性系统结构辨识问题转化而成。仿真实验表明,遗传算法求解非线性系统辨识比其他方法具有更好的近似解,证明了该算法的有效性和实用性。For identification and controlling study of linear, time-invariant object with uncertain pa- rameters, adaptive scheduling rules and structure optimization of the function composition of identify and controllers in these studies, aiming to ensure the identification systems globally stable, are built on top of linear system theory. Nonlinear system identification and adaptive control hare been difficult to find the appropriate mathematical methods, presenting a mathematical model for nonlinear system, constituting a new method of system identification, and combinatorial optimization problems by transformation and struc- tural identification of nonlinear system are transformed by problems of structural identification of nonlinear system. Computer simulation experiments show that genetic algorithms for solving nonlinear systems identi- fication has better approximate solutions compared to other methods, which provides effectiveness and practicality of the algorithm＇ s.

关 键 词：系统辨别 遗传算法 非线性 径向基函数 GMDH方法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]