Using fuzzy sets in determining the class of the car

Abstract

The article presents the stages of constructing a fuzzy inference system for selecting a class of a car. The procedures for collecting and processing the expert information necessary for constructing a system of fuzzy inference are described. Linguistic variables are chosen, fuzzy variables "car state", "car class", a clear variable "car category" are constructed, and the intervals for the variable "year", necessary for constructing a fuzzy model, are chosen based on expert data. Productive rules for the system of fuzzy inference are formed. The proposed procedures and methods were implemented as a system of fuzzy in the MatLab environment. Stages of construction and analysis of the adequacy of the fuzzy model are presented. The graphical interface of the variable editor, rule editor, and the fuzzy output surface of the model, developed in the MatLab environment, is presented. The received model allows to establish dependence of values of an output variable «a class of the car» from values of input variables «a category of the car», «year», «a condition of the car». The simulation results automatically change when the parameters of input variables change, which allows using this model under changing external conditions. The results obtained in the modeling process are further used to construct the model for choosing the optimal trip, and the variable "car class" becomes the input one.