When Will a Genetic Algorithm Outperform Hill Climbing
Melanie S. Mitchell, John H. Holland, Stephanie Forrest · PDXScholar (Portland State University) · 1993
In this paper we review some previously published experimental results in which a simple hillclimbing algorithm-Random Mutation Hill-Climbing (RMHC)-significantly outperforms a genetic algorithm on a simple "Royal Road" function. vVe present an analysis of RMHC followed by an analysis of an "idealized" genetic algorithm (IGA) that is in turn significantly faster than RMHC. We isolate the features of the IGA that allow for this speedup, and discuss how these features can be incorporated into a real GA and a fitness landscape, making the GA better approximate the IGA. We use these features to design a modified version of the previously published experiments, and give new experimental results comparing the GA and RMHC.