The Breeder Genetic Algorithm (BGA) was designed according to the theories and methods used in the science of livestock breeding. The prediction of a breeding experiment is based on the response to selection (RS) equation. This equation relates the change in a population's fitness to the standard deviation of its fitness, as well as to the parameters selection intensity and realized heritability. In order to maximize the response, the product of the standard deviation and the heritability has to be maximized. This criterion has been used to design new recombination operators. In this talk we will apply this framework to populations having a spatial structure. We will discuss the following two models: a population living on separate islands with occasional migration, and an ecology consisting of different subspecies which compete against each other. For both models we will derive difference equations which describe the evolution of the population. By designing the subspecies to use different search strategies, the competition model is able to realize a multi-resolution search. Numerical results for continuous multi-modal functions will be presented.

Back to Workshop Schedule

1996-1997 Mathematics in High Performance Computing