Evolutionary modeling algorithm of ordinary differential equations based on genetic modeling

Abstract

Differential equations are often used to describe complex systems and nonlinear systems related to time, but it is difficult to establish an ideal model for such systems based on some observed data. Especially when dealing with unknown chaotic system data, it is blind and difficult to combine existing nonlinear analysis results with relevant experience. In this paper, through the study of genetic modeling, the optimization process of model parameters using GA is embedded in the optimization process of model structure using GP, and the local search process of neighborhood solution generated by GP-based standard mutation operator is carried out for some individuals in each evolution generation, and the evolution modeling algorithm of ordinary differential equations is designed and implemented, and an application example is given.