Parameter Control within a Co-operative Co-evolutionary Genetic Algorithm

Typically GAs have a number of fixed control parameters which have a significant effect upon the performance of the search. This paper deals with the effects of self-adapting control parameters, and the adaptation of population size within the sub-populations of a coevolutionary model. We address the need to investigate the potential of these adaptive techniques within a co-evolutionary GA, and propose a number of model variants implementing adaptation. These models were tested on some well known function optimisation problems. The experimental results show that one or more of the model variants yield improvements over the baseline co-evolutionary model.