Effects of Isolation in a Distributed Population Genetic Algorithm

An investigation is reported into the effects of isolation within a population upon its evolution under a genetic algorithm. The effects of varying population size were compared between a panmictic and a continuously distributed population, each offered an identical environment. Theoretical analysis predicted a logarithmic variation of performance with population size alone. Experiment confirmed such a variation in both populations. Although the distributed population always obtained superior peak fitness, no improvement in the ability to discover optima could be attributed to isolation. However, the distributed population was clearly observed to better maintain allelic diversity. Evidence was also found for the emergence of meta-stable geographical boundaries. Possible explanations are offered.

[1]  David E. Goldberg,et al.  Genetic Algorithms and the Variance of Fitness , 1991, Complex Syst..

[2]  Kenneth A. De Jong,et al.  Genetic Algorithms are NOT Function Optimizers , 1992, FOGA.

[3]  Dana S. Richards,et al.  Punctuated Equilibria: A Parallel Genetic Algorithm , 1987, ICGA.

[4]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[5]  Reiko Tanese,et al.  Parallel Genetic Algorithms for a Hypercube , 1987, ICGA.

[6]  Ian East,et al.  An investigation of several parallel genetic algorithms , 1991 .

[7]  John J. Grefenstette,et al.  A Parallel Genetic Algorithm , 1987, ICGA.

[8]  David R. Jefferson,et al.  Selection in Massively Parallel Genetic Algorithms , 1991, ICGA.

[9]  Lashon B. Booker,et al.  Proceedings of the fourth international conference on Genetic algorithms , 1991 .

[10]  John J. Grefenstette,et al.  Deception Considered Harmful , 1992, FOGA.

[11]  D. E. Goldberg,et al.  Simple Genetic Algorithms and the Minimal, Deceptive Problem , 1987 .

[12]  Melanie Mitchell,et al.  What makes a problem hard for a genetic algorithm? Some anomalous results and their explanation , 1993, Machine Learning.

[13]  David E. Goldberg,et al.  Sizing Populations for Serial and Parallel Genetic Algorithms , 1989, ICGA.

[14]  D. Wolpert,et al.  No Free Lunch Theorems for Search , 1995 .

[15]  Jon Rowe,et al.  Direct Replacement: A Genetic Algorithm Without Mutation Which Avoids Deception , 1993, Evo Workshops.

[16]  Martina Gorges-Schleuter,et al.  ASPARAGOS An Asynchronous Parallel Genetic Optimization Strategy , 1989, ICGA.

[17]  Paul Bryant Grosso,et al.  Computer Simulations of Genetic Adaptation: Parallel Subcomponent Interaction in a Multilocus Model , 1985 .

[18]  J. David Schaffer,et al.  Proceedings of the third international conference on Genetic algorithms , 1989 .

[19]  Lawrence Davis,et al.  Genetic Algorithms and Simulated Annealing , 1987 .