Usage of imperialist competitive optimization algorithm (ICA) for dynamic systems modelling with ANFIS

Imperialist Competitive Algorithm is a social based heruristic optimization algorithm that deals with the challenge of some countries with certain populations among themselves so as to be the emperial power. In this study, ICA's (Imperialist Competİtive Algorithm) optimization performance over the problem of non-linear dynamic system modelling has been studied using ANFIS (Adaptive Nuro Fuzzy Inference System) fuzzy model. For the study, sample dynamic systems selected from the literature have been used. Parameters of each sample dynamic system have been modelled with ANFIS preferred due to its flexible and mathematical structure using ICA optimization algorithm. For each system, the outcomes obtained with ICA (Standard ICA, Std-ICA) have been compared to the outcomes of the two new algorithms developed from ICA during this study. According to the obtained statistical data, it has been observed that performance of G1-ICA algorithm was better than that of the others to two ICA algorithms. Finally, performance of ANFIS models obtained in this study have been statistical compared with the ANFIS models optimized using the other heuristic algorithms. According to this review, it has been observed that ICA revealed values close to the desired.