GENERATION AND CLASSIFICATION OF STRUCTURAL TOPOLOGIES WITH GENETIC ALGORITHM SPECIATION

Extending previous efforts, this article describes how a speciating ge netic algorithm is used to distribute subsets of the evolving populati on of solutions over the design space. This distribution of solutions is analogous to different species exploiting different niches in an ec osystem. In addition to reviewing genetic algorithms with an emphasis on techniques to cause such niche exploitation, we describe how we use statistical cluster analysis techniques to quantify the extent to whi ch a population is speciated and how this measure can be used to proba bilistically encourage mating of reasonably similar designs (i.e., int raspecies mating ). Results demonstrate the creation of different good designs of characteristically different topology and shape.