基于遗传算法的时变时滞参数辨识  

On-line Identification of Time-varying Time-delay Based on Genetic Algorithm for Large Time-delay Systems

作　　者：潘峰[1] 冯冬梅[1] 韩如成[1] 

机构地区：[1]太原重型机械学院电子信息工程分院,太原030024

出　　处：《仪器仪表学报》2004年第z1期910-912,918,共4页Chinese Journal of Scientific Instrument

基　　金：山西省青年科技研究基金资助(20031030)。

摘　　要：针对时滞系统的时变时滞提出一种新的在线辨识方法。该方法基于采样到的输入输出数据,采用遗传算法对时变时滞进行跟踪,它可在比较大的参数空间内迅速寻优,尽可能获得全局最优解。即它能更准确地辨识出时滞参数,以便于预估补偿,从而减少时滞引起的影响。仿真结果表明该辨识方法的精度是比较高的。It presents a new approach to on-line identification of time-varying for time-delay system. Based on sampled input-output data, the GA (genetic algorithm) is used to track the time-varying time-delay, which can simultaneously evaluate many points in the parameter space, and increase the likelihood of finding the global optimum. Then time-delay that is applied to predictive mode can been identified more exactly, which can suppress the effect of time-varying time-delay. The simulation results show that the identification precision is proved through using the approach.

关 键 词：大时滞系统 遗传算法 系统辨识 

分 类 号：TH7-55[机械工程—仪器科学与技术]