作者: J. W. Duda , M. J. Jakiela
DOI: 10.1115/1.2828774
关键词: Measure (mathematics) 、 Genetic algorithm 、 Data mining 、 Ecological niche 、 Niche 、 Computer science 、 Artificial intelligence 、 Genetic algorithm 、 Network topology 、 Population 、 Topology (chemistry)
摘要: Extending previous efforts, this article describes how a speciating genetic algorithm is used to distribute subsets of the evolving population solutions over design space. This distribution analogous different species exploiting niches in an ecosystem. In addition reviewing algorithms with emphasis on techniques cause such niche exploitation, we describe use statistical cluster analysis quantify extent which speciated and measure can be probabilistically encourage mating reasonably similar designs (i.e., intraspecies mating). Results demonstrate creation good characteristically topology shape.