Multi-layered Niche Formation
Recently an abstraction of genetic algorithms has been developed in which a population of GAs in any epoch is represented by a single vector whose elements are the the probabilities of the corresponding bit positions being equivalent to 1. The process of evolution is represented by learning the elements of the probability vector. We have previously extended this to model homeotic genes which are environmentally driven and turn other genes on and off. In this paper we incrementally develop the algorithm on a set of standard problems used to compare methods for the simultaneous optimisation of conflicting criteria within a single population.
KeywordsGenetic Algorithm Learning Rate Probability Vector Realisable Gene Homeotic Gene
Unable to display preview. Download preview PDF.
- S. Baluja and R. Caruana. Removing the genetics from the standard genetic algorithm. In Proceedings of the Twelfth International Conference on Machine Learning, 1995.Google Scholar
- J. Cohen and I. Stewart. The Collapse of Chaos. Penguin, 1995.Google Scholar
- C. Fyfe. Developing and understanding niche formation in dynamically-changing environments using structured evolutionary algorithms. In International Symposia on Soft Computing and Intelligent Industrial Automation, 1996.Google Scholar
- T. Kohonen. Self-Organising Maps. Springer, 1995.Google Scholar
- J.M. Mendel. A Prelude to Neural Networks: Adaptive and Learning Systems. Prentice Hall, 1994. (Ed).Google Scholar
- C. Ryan. Racial harmony and function optimization in genetic algorithms-the races genetic algorithm. In Proceedings of EP95. MIT Press, 1995.Google Scholar