细菌生存优化在非线性模型辨识中的应用  

Application of Bacterial Foraging Optimization in Non-linear Model Identification

作　　者：林卫星[1] Peter X.Liu 李文磊[1] 陈炎海[1] 欧超[1] 

机构地区：[1]宁波大学信息科学与工程学院,宁波315211 [2]Department of Systems and Computer Engineering,Carleton University,Ottawa,ON,K 1S 5B6

出　　处：《系统仿真学报》2009年第10期3100-3104,共5页Journal of System Simulation

基　　金：浙江省自然科学基金项目(Y107010);宁波市自然科学基金项目(2008A610019)和(2008A610025)

摘　　要：提出了一种新的基于细菌生存优化(Bacterial Foraging Optimization BFO)的非线性模型辨识方法。它是利用群集智能仿生BFO算法对一类Hammerstein系统进行辨识,从而估计出它的参数模型。通过对这类输入非线性模型进行辨识,并用仿真实验说明BFO算法的参数设置与选择方法。比较基于粒子群优化(Particle Swarm Optimization PSO)的非线性模型辨识算法,特别是对有色噪声的鲁棒性、模型的辨识精度、辨识收敛速度进行对比分析,以得出BFO辨识算法的优缺点及其有效性。A new identification algorithm of the non-linear system was proposed based on bacterial foraging optimization （BFO）. In swarms intelligence, the biomimicry of BFO algorithm was used to identify a Hammerstein model and its parameters. The validity and the influence of BFO algorithm＇s selection of parameters and identification results by the experiment are showed. BFO identification algorithm was compared with particle swarm optimization （PSO） algorithm, and its robustness of colored noise, the identification precision of the model, the convergence of parameters and so on were analyzed.

关 键 词：系统辨识 HAMMERSTEIN模型 细菌生存优化 粒子群优化 

分 类 号：TP271[自动化与计算机技术—检测技术与自动化装置]