A comparative study of genetic operators for controller parameter optimisation

The objective of this paper is to propose a Genetic Algorithm (GA) scheme that works well in a spectrum of controller parameter optimisation problems. A set of 18 different GAs, which combine 3 different selection methods, 3 probabilities of crossover and 2 probabilities of mutation, are used to solve 4 controller parameter optimisation problems and the results obtained are compared. In each of these problems the control methodology varies so the study covers a variety of current control research areas.