基于改进差分进化算法的Wiener模型辨识  

Identification of Wiener Model Based on Improved Differential Evolution(SADE) Algorithm

作　　者：徐小平[1] 白博[1] 钱富才[2] 

机构地区：[1]西安理工大学理学院,西安710054 [2]西安理工大学自动化与信息工程学院,西安710048

出　　处：《系统仿真学报》2016年第1期147-153,共7页Journal of System Simulation

基　　金：国家自然科学基金(61273127);陕西省教育厅专项科研计划项目(14JK1538);陕西省自然科学基础研究计划项目(2014JM8325)

摘　　要：针对非线性Wiener模型的参数辨识问题,提出了一种基于Sigmoid函数及自适应算子改进差分进化(improved differential evolution algorithm with Sigmoid function and adaptive mutation operator,SADE)算法的参数辨识方法。利用Sigmoid函数及自适应变异算子改进了基本差分进化算法的变异操作部分,改进的方法能够有效地克服基本差分进化算法的过早收敛和不稳定性等缺点。将该改进差分进化算法应用于对非线性Wiener模型的参数辨识问题,达到了较高的辨识精度。在仿真试验中,与其它已有方法进行比较,仿真结果说明了所给的参数辨识方法是合理和有效的。To identify the parameters of nonlinear Wiener model, a new identification method was put forward based on a SADE algorithm. Its basic idea is as follows： the Sigmoid function and adaptive mutation operator were adopted to improve the mutation operation part of the basic differential evolution algorithm, accordingly, the disadvantages of the basic differential evolution algorithm, such as premature convergence, instability, etc, were effectively overcome. The proposed algorithm was used to parameter identification problem of Wiener model; moreover, the accuracy of identification was well improved. In the numerical simulation, compared with other relevant existing algorithms, simulation results show that the proposed method is rational and effective.

关 键 词：差分进化算法 SIGMOID函数 自适应算子 SADE算法 WIENER模型 辨识 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]