The effect of population enrichment in genetic programming

The paper examines the effect of “population enrichment” in genetic programming as a means of efficiently discovering promising directions for solution exploration in a large problem space. With genetic programming it is advantageous to not restrict the size or shape of the solution and enrichment offers an efficient way to present the initial population with interesting options for development. Multiple abbreviated runs were made, using different random seeds, to keep the size of the members small and save processing time. The best member from each abbreviated run was used to create an enriched population which was loaded along with a full complement of randomly generated unique members at the beginning of a consolidated run