Identification of Nonlinear Discrete Dynamic Systems Using Enhanced Genetic Programming

This paper introduces the use of the evolutionary genetic programming to solve the identification problem of nonlinear discrete dynamic systems. The standard GP is enhanced by a general parameter optimization technique to include one of four proposed techniques. The efficiency of finding the numeric constant node is significantly improved as compared to the traditional methods. The simulation procedure includes first a comparison between the enhanced GP by one of the parameter optimization techniques and the standard GP. Then after for six different models, a comparison between the four utilized techniques is performed. The comparison is made in terms of the number of runs require to find the perfect model, and the average number of generations for successful runs.