Research On Identification Of Nonlinear System Based On Improved Differential Evolution Algorithm

There are a large number of non-linear links in the actual production process,and these processes are often accompanied by the interference of heavy tail noise,which seriously affects the accuracy of system identification.Research shows that the fractional-order model contains more information than the integer-order model and is closer to the real process.Therefore,studying the identification of fractional-order nonlinear systems and the identification of nonlinear systems under heavy-tailed noise has important theoretical and application value.This article proposes corresponding identification methods for the above problems,as follows:1.An adaptive neighborhood search differential evolution algorithm(ADELS)is proposed.The robustness of the algorithm is improved by introducing Chebyshev sequence.The operator is updated adaptively according to the value of the individual fitness function,making its local search and global search capabilities It has been improved,and the method of sub-population is adopted in the mutation operation to further strengthen the optimization ability of the algorithm.The simulation experiment is compared with other intelligent optimization algorithms based on test functions,which shows the efficiency of the ADELS algorithm.2.Aiming at the influence of heavy-tailed noise on system identification,this paper considers the addition of mixed Gaussian noise in the simulation experiment.The ADELS algorithm is proposed in this paper.The ADELS algorithm and other intelligent optimization algorithms are used to identify the Hammerstein model under Gaussian mixed noise.The simulation results prove the efficiency of the ADELS algorithm.3.In the actual production process,the fractional-order nonlinear model,as an extension of the integer-order model,has more system internal information and has important research value.This paper uses the ADELS algorithm to estimate the initial values of the parameters of the fractional Hammerstein model.Based on the theory of fractional calculus,the recursive identification algorithm of the fractional Hammerstein model is derived,combined with the initial values of the parameters provided by the ADELS algorithm,together to complete the identification of the parameters of the fractional Hammerstein model,including the coefficients and scores of the nonlinear part and the linear part the order is obtained through algorithm identification.The simulation results show that the performance of the proposed identification method is better.