Aspects of adaptation in natural and artificial evolution

This work addresses selected aspects of natural evolution, especially of the field of population genetics, that are considered to be meaningful for algorithmic further developments in the field of Genetic Algorithms (GAs) and Genetic Programming (GP) by the authors. In this connection special attention is devoted to selection and replacement strategies, as these are exactly the aspects that do not depend on certain problem representations and corresponding operators and therefore allow generic algorithmic further development. The concept of offspring selection is described as an example of such a problem independent further developed algorithmic concept, which allows to maintain the relevant genetic information stored in a population more efficiently. The potential of this new selection strategy is pointed out in terms of references to recent results achieved on the basis of well known benchmark problems in the field of GAs and GP. Categories and Subject Descriptors I.2.8 [Artificial ...

• Type: Conference
• Year: 2007
• Where: GECCO
• Authors: Michael Affenzeller, Stefan Wagner 0002, Stephan M. Winkler
• Keywords: Algorithmic, GECCO 2007, Optimization, Population Genetics, Selection