Genetic algorithm tuning of a fuzzy logic controller for a dynamic system

This thesis describes the use of the genetic algorithm to facilitate the design process of a fuzzy logic based controller. The basic mechanics of both the fuzzy logic control system and the genetic algorithm are presented for establishing the foundation of this research. A detailed design process to integrate the fuzzy logic controller and the genetic algorithm is disclosed. Software is developed to simulate the dynamics of a two-link planar manipulator, and a multiple-link fuzzy logic control (MLFLC) system is developed to control a non-linear robotics system. In this work, the genetic algorithm technique is used to design both the Universe of Discourse of some of the control variables and also the shape and location of membership functions. As a result, the system is able to maintain control with moderate errors. The most important contribution of this work is that it demonstrates the effectiveness of using genetic algorithms to optimize fuzzy logic control systems.