Supervision of fuzzy controllers using genetic algorithms

The fast growth of fuzzy control applications in industry, together with the need of getting increasingly efficient control systems, has been the motivation for the development of approaches to the supervision of fuzzy controllers. Some novel approaches to this problem encompass the application of artificial intelligence techniques, such as the utilization of neural networks and genetic algorithms. In this article an approach based on the genetic algorithms (GAs) technique is proposed, allowing the simultaneous determination of: (i) the set of fuzzy rules, (ii) the shape of membership functions, and (iii) the universes of discourse for the linguistic variables. To illustrate the application of the suggested approach, two systems were selected: an inverted pendulum and a manufacturing system