ABSTRACT
The concept of offspring selection as described in Chapter 4 is very well suited to be transferred to parallel GA concepts. In the sense of parallel GA nomenclature, our proposed variant called SASEGASA (which stands for self-adaptive segregative genetic algorithm with simulated annealing aspects) is most closely related to the class of coarse-grained parallel GAs. The well-known island model supports global search by taking advantage of the steady pulsating interplay between breadth search and depth search supported by the forces of genetic drift and migration.
