The Influence of Elitism Strategy on Migration Intervals of a Distributed Genetic Algorithm
A distributed genetic algorithm is an important technique on practical use. To parallelize distributed genetic algorithms, researchers have been discussing various advanced algorithms with reduced migrations. An interesting previous study shows that both too long migration interval and too short migration interval cause a degraded performance in finding solutions. This paper assumes that a cause degrading the performance is a behavior of elites and also discusses a modified elitist model. The experiments show that a modified elitist model improves the performance even if the migration intervals are not set appropriately. These results seem to be design guides for discussing distributed genetic algorithms with reduced migrations.
Keywordsevolutionary computation genetic algorithm island model multiple populations migration elitist model convergence
Unable to display preview. Download preview PDF.
- 1.Goldberg, D.E.: Genetic Algorithm in Search, Optimization and Machine Learning. Addison Wesley Longman Publishing, Boston (1989)Google Scholar
- 3.Tanese, R.: Distributed Genetic Algorithms. In: Schaffer, J.D. (ed.) Proceedings of the 3rd International Conference on Genetic Algorithms, pp. 434–439. Morgan Kaufmann Publishers, Virginia (1989)Google Scholar
- 5.Skolicki, Z., De Jong, K.A.: The influence of migration sizes and intervals on island models. In: Beyer, H., O’Reilly, U. (eds.) Proceedings of the 2005 Conference on Genetic and Evolutionary Computation, GECCO 2005, pp. 1295–1302. ACM, Washington DC (2005)Google Scholar
- 6.Skolicki, Z.: An analysis of island models in evolutionary computation. In: Beyer, H., O’Reilly, U. (eds.) Proceedings of the 2005 Conference on Genetic and Evolutionary Computation, GECCO 2005, pp. 386–389. ACM, Washington DC (2005)Google Scholar
- 10.De Jong, K.A.: An Analysis of the Behavior of a Class of Genetic Adaptive Systems. Ph.D Thesis, University of Michigan (1975)Google Scholar
- 11.Hiroyasu, T., Miki, M., Negami, M.: Distributed Genetic Algorithms with Randomized Migration Rate. In: Proceedings of 1999 IEEE International Conference on Systems, Man, and Cybernetics Conference, vol. 1, pp. 689–694. IEEE, Tokyo (1999)Google Scholar
- 12.Munetomo, M., Takai, Y., Sato, Y.: An Efficient String Exchange Algorithm for a Subpopulation-Based Asynchronously Parallel Genetic Algorithm and Its Evaluation. Transactions of IPSJ 35(9), 1815–1827 (1994) (in Japanese)Google Scholar