Nonlinear System's Identification using Neuro-Fuzzy model tuned by Crow Search Algorithm

Abstract

We propose in this work a new algorithm of optimization named Crow Search Algorithm CSA to elicit neuro-fuzzy model such TS type. In the proposed study, a particle is formed by two tasks: its structure and its parameters. The CSA algorithm was compared with others: GA and PSO through a modeling of nonlinear system. The results prove that CSA method gives optimal mean of MSE and optimal of standard deviation of MSE compared to GA and PSO.