Fuzzy Controller Design by Ant Colony Optimization

Abstract

An ant colony optimization (ACO) application to a fuzzy controller design, called ACO-FC, is proposed in this paper for improving design efficiency. A fuzzy controller's antecedent part, i.e., the "if" part of its composing fuzzy if-then rules, is partitioned in grid-type, and all candidate rule consequent values are then listed. An ant tour is regarded as a combination of consequent values selected from every rule. A pheromone matrix among all candidate consequent values is constructed. Searching for the best one among all combinations of rule consequent values is based mainly on the pheromone matrix. The proposed ACO-FC performance is shown to be better than other evolutionary design methods on one simulation example.