SEARCH OF OPTIMAL MEMBERSHIP FUNCTIONS OF FUZZY SYSTEMS BASED ON BIOINSPIRED EVOLUTIONARY ALGORITHMS. PART II. METHOD IMPLEMENTATION AND STUDY OF ITS EFFICIENCY

Abstract

This article is devoted to the efficiency of a method of optimal membership functions search for fuzzy systems based on bioinspired evolutionary algorithms of global optimization. The proposed method allows finding the optimal membership functions of linguistic terms at solving the compromise problem of minimizing the objective function and reducing computational costs in the process of further parametric optimization of fuzzy systems. To study the effectiveness of the considered method in this work, the search of the optimal membership functions is conducted for a fuzzy controller of the control system of a multi-purpose mobile robot designed to move along inclined and vertical ferromagnetic surfaces, with the implementation of this method based on 3 bioinspired evolutionary algorithms: genetic, artificial immune systems, biogeo­graphic. The analysis of the obtained results of computer modeling showed that the usage of the proposed method of search of optimal membership functions gives the opportunity to increase significantly the efficiency of the mobile robot control, as well as to reduce the total number of parameters at further parametric optimization of linguistic terms, which confirms the high efficiency of the developed method.