Proceedings of the 2008 GECCO conference companion on Genetic and evolutionary computation - GECCO '08
Full text: Download
In this paper we describe an improvement of an entropy-based diversity preservation approach for evolutionary algorithms. This approach exploits the information contained not only in the parts that compose an individual, but also in their position and relative order. We executed a set of preliminary experiments in order to test the new approach, using two different problems in which diversity preservation plays a major role in obtaining good solutions.