The Nature of Niching: Genetic Algorithms and the Evolution of Optimal, Cooperative Populations

242 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1997.We extend the above analysis by looking at some aspects of the general case of multiple overlapping niches. We discover that calculating the equilibrium point for three or more niches under resource sharing can be computationally expensive, as evidenced by the difficulty that GA selection has in reaching it. We present some evidence that equilibrium can thus represent solutions to hard problems, and that selection plus sharing might be a computationally intensive yet efficient algorithm (even without the exploration operators of recombination and mutation). (Abstract shortened by UMI.).U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD