 |
    |
Demonstrating the Evolution of Complex Genetic Representations: An Evolution of Artificial Plants
Marc Toussaint
Institut für Neuroinformatik,
Chair for Theoretical Biology
Ruhr-Universität
Bochum ND-04, 44780 Bochum,
Germany
toussaint@neuroinformatik.rub.de
Abstract. A common idea is that complex evolutionary adaptation is enabled by complex genetic representations of phenotypic traits. This pa-per demonstrates how, according to a recently developed theory, genetic representations can self-adapt in favor of evolvability, i.e., the chance of adaptive mutations. The key for the adaptability of genetic representations is neutrality inherent in non-trivial genotype-phenotype mappings and neutral mutations that allow for transitions between genetic representations of the same phenotype. We model an evolution of artificial plants, encoded by grammar-like genotypes, to demonstrate this theory.
LNCS 2723, p. 86 ff.
lncs@springer.de
© Springer-Verlag Berlin Heidelberg 2003