CIS Logo

University of Oregon
Computer & Information Science

Artificial Evolution

Principal members are:

Art Farley

Research in artificial evolution reduces the world to binary strings and procedures operating on binary strings. There is a rich history of research in artificial evolution adopting either theoretical or experimental approaches. Theoretical approaches view evolution as searching through abstract mathematical landscapes defined by fitness functions and establish theorems regarding how search proceeds in these differing landscapes. Experimental approaches to artificial evolution select particular representations for genotypes, forms of fitness functions, and values for probabilistic parameters and generates instances of the resultant population dynamics through simulation. We apply simulations, based in genetic algorithms, to the investigation of our research questions.

We have considered the issue of populations structure and how it impacts the dynamics of evolution, as measured in term of a population's average fitness and genetic diversity. Population structure is any factor other than fitness that impacts the selection of parents in generating a successive generation. We model an arbitrary population structure as a graph and imbed individuals of a population in the graph. An individual is restricted to mating with individuals located at neighboring vertices of the graph.

We have considered the impact of developmental learning on evolution, where developmental learning is represented as a development process that interacts with the environment and chooses better expression of available genes based upon that interaction. In a related study, we have investigated the role of dominance in diploid evolution. Dominance is the preferential expression of certain gene alleles at heterozygotes of diploid genotypes.


Farley, A.M., "Population Structure and Artificial Evolution", Proceedings of the Senventh International Conference on Artificial Evolution, Berlin:Springer-Verlag, LNCS 3871,p 213-225, (2006).

Farley, A.M., "Development and choice", Proceedings Genetic and Evolutionary Computation Conference (GECCO07) , London, England, Late Breaking Papers (2007).

Farley, A.M., "Developmental Learning in Changing Environments", in preparation.

send feedback to: