Finite Markov Chain Models of an Alternative Selection Strategy for the Genetic Algorithm

T his paper present s finite Markov chain models of the select ion strategy known as Boltzmann tournament select ion. Unlike previous research at the string level, this st udy represents popul at ions at the more general, equivalence-class level. The changing distribution of classes is analyzed using Markov chains , and a Markov chain model is used to predict expected dr ift t ime for the select ion procedure. The accuracy of the final model is verified through direct simulation.