A Study of Mate Selection Schemes in Genetic Algorithms—Part I

In the Genetic Algorithm (GA) literature, many models focus on problems where each individual’s fitness is independent of others (or implicitly defined by others). In this paper, a framework is introduced for studying mate selection in the context of GA. The objective is to model interdependent fitnesses of population individuals by allowing them to search for mates. The resulting GA thus forms a more complex system in which each individual’s fitness depends on both the environment and other population members. The methods for investigation consist of the Schema Theorem, a Markov chain analysis, along with several empirical results. I will show that mate selection plays a crucial role in GA’s search power. In particular, individuals with more distinct characteristics collectively facilitate the search for a single, better solution. The results presented on the effects of mate selection are a first step toward a deeper understanding of how GAs work, and thus how to design more robust GAs.