The Use of Genetic Algorithms for Nonlinear Controller Design

Abstract

The optimization of nonlinear controller parameters by Genetic Algorithm (GA) is explored in this paper. The type of nonlinear controller that is considered is derived from sliding mode control theory, which is known for its robust properties. The GA technique for optimization has developed from its foundation in Darwinian evolution into a powerful optimization algorithm that can be used for parametric design. Both sliding mode control theory and GA theory are presented. The use of an elite GA for sliding mode controller parameter optimization is illustrated by means of an example which involves the design of a heading controller for a scale model of a supply ship. A GA optimized controller solution is presented. The performance of this controller is illustrated through simulation studies and trials in a model ship water tank facility. These evaluations illustrate that both the sliding mode control law and GA technique are effective design methods for automated design for nonlinear controllers.