Control of industrial processes using rules and membership functions adjustment by genetic algorithms

Resumen

The purpose of this paper is to present a strategy for the control of industrial processes through the automatic adjustment of the rules and of the membership using Genetic Algorithms. For such, it was developed and implemented an algorithm that transforms the rules and the membership functions in chromosomes that are submitted to an evolution, crossover and mutation. The general idea is to obtain a new family of rules and membership functions that better can control a process, optimizing the final result. The proposed algorithm was previously incorporate to the Computational Package for the Teaching of the Fuzzy Logic developed that he/she has the objective of teaching the fuzzy logic for control students. This package contains all of the necessary instructions for the users to understand the beginnings of the diffuse control. The main objective of the package is to park a vehicle, leaving of any initial position, in a garage. To accomplish this task, the students initially should develop a group of rules of fuzzy control and membership functions that will define the path of the vehicle. The fuzzification and defuzzification processes of the variables are accomplished by the program without the user's interference.