Summary: Evolving Better Representations through
Selective Genome Growth
Lee Altenberg
Institute of Statistics and Decision Sciences, Duke University,
Durham, NC 27708­0251 U.S.A. Internet: altenber@acpub.duke.edu
Abstract---
The choice of how to represent the search space for
a genetic algorithm (GA) is critical to the GA's per­
formance. Representations are usually engineered by
hand and fixed for the duration of the GA run. Here a
new method is described in which the degrees of free­
dom of the representation --- i.e. the genes -- are in­
creased incrementally. The phenotypic effects of the
new genes are randomly drawn from a space of differ­
ent functional effects. Only those genes that initially
increase fitness are kept. The genotype­phenotype map
that results from this selection during the construction
of the genome allows better adaptation. This effect is
illustrated with the NK landscape model. The resulting
genotype­phenotype maps are much less epistatic than

Source: Altenberg, Lee - Department of Information and Computer Science, University of Hawai'i at Manoa

Collections: Computer Technologies and Information Sciences