Genetic Algorithms in Search, Optimization and Machine Learning
Genetic Algorithms in Search, Optimization and Machine Learning
An Investigation of Niche and Species Formation in Genetic Function Optimization
Proceedings of the 3rd International Conference on Genetic Algorithms
Radical epistasis and the genotype-phenotype relationship
Artificial Life
| Hi-index | 0.00 |
A genetic system that uses a radically epistatic mapping from genotype to phenotype is defined. This system is proposed in order to study the population dynamics rather than create a new system for function optimization. Two behaviors not found in traditional genetic systems are explored: first, although the genetic system uses only simple generational recombination and fitness scaling, it is capable of distributing a population of solutions on a number of non-adjacent phenotype peaks. This capability is analyzed and empirically explored. Second, the genetic system is shown to have the capability of tracking an environment where the adaptive peaks are changed in a recurrent manner.