System identification using dynamic ga based on numeric encoding

Abstract

A dynamic genetic algorithms based on numeric encoding is proposed and its application in system identification is discussed. Simulation shows that the introduction of both numeric encoding and dynamic mutation can effectively improve the accuracy and speed of searching for the optimum. It also show that the improved Genetic algorithm can identify time delay and parameters of the plant at the same time and converge to globle optimization.

Keywords

genetic algorithms / dynamic mutation operator / system identification

References

[1] KrisinssonDumout. System identification and control using genetic algorithms. IEEE Trans on Syst Man and Cybem, 1992, 22(5): 1033-1046

[2] HuangJiong, WuYongge, LiGajun, et al. . On-line system identification based on genetic algorithm. Imformation and Control, 1996, 25(3): 171-175(in Chinese)

[3] JeongIL-Kwon, LeeJujang. Adaptive simulated annealing genetic algorithms for control application. International Journal of System Science, 1996, 27(2): 241-253

[4] Patnaik LalitM. Genetic algorithms: A survey. Computer, 1994, 27(6): 17-26

[5] SuDiqianAdaptive control, 1993, Shengyan, Dongbei University Press: 2428(in Chinese)