Reducing disruption of superior building blocks in genetic algorithms
Arthur L. Corcoran, Roger L. Wainwright · 1996
The strength of genetic algorithms is the ability to recombine short, highly-fit schemata (i.e., building blocks) to produce longer, more highly-fit schemata.Unfortunately, this benefit can be diminished by the effects of hitchhiking, allele loss, disruption, etc.In many cases this is due to the absence of information to indicate the relative strengths of the genes or the bonding between the genes.Consequently, beneficial building blocks are just as likely to be disrupted as harmful ones.In this paper, we describe several methods which can be used to preserve sequences and distributions of beneficial building blocks.Most research to date attempts to minimize the negative aspects of disruption.This research emphasizes the positive aspect Of identifying and manipulating optimal sequences.