Co-evolutionary genetic optimization based ordinary differential equations identification for multi-input multi-output chaotic systems

A Co-evolutionary algorithm based ordinary differential equations of multi input multi output continue chaos system identification is proposed. Structures of ODEs are optimized using GP (Genetic Programming), and corresponding parameters of ODEs are tuning via GA. Fitness degrees of these individuals which represent suitable structure of ODEs are high than others, because their parameters are optimized by GA. Both structures and corresponding parameters will be evolved using the proposed coevolution GP-GA method. Population initialization, genetic operator, fitness calculation and evolution scheme of multi-population GP are given. Comparisons are made among the proposed method, RBF neural network based method and fuzzy clustering based method, simulation results show the efficiency of the method.