ABSTRACT
The ultimate goal of the extended algorithmic concepts described in this chapter is to support crossover-based evolutionary algorithms, i.e., evolutionary algorithms that are ideally designed to function as building-block assembling machines, in their intention to combine those parts of the chromosomes that define high quality solutions. In this context we concentrate on selection and replacement which are the parts of the algorithm that are independent of the problem representation and the according operators. Thus, the application domain of the new algorithms is very wide; in fact, offspring selection and the RAPGA (a special variant of adaptive population sizing GA) can be applied to any application that can be treated by genetic algorithms (of course also including genetic programming).
