Sampling Methods in Evolutionary Computation: Assuring Genetic Diversity and Stochastic Selection

JORGE E. NÚÑEZ MC LEOD

Evolutionary Algorithms and Intelligent Tools in Engineering Optimization

CIMNE-WIT PRESS UK

Lugar: Barcelona-Southampton; Año: 2005; p. 52 - 71

Abstract. This chapter focuses on sampling methods with emphasis in stochastic sampling methods. The chapter begins with a brief review about how sampling methods impact in Evolutionary Computation (EC) performance. The next sections develop the theory and the results about sampling stochastic generation methods. The answers of several questions about EC performance are behind these methods. The explanation attempts clarify those issues related to genetic diversity or to gene or individual selections. The third section describe the specific mechanisms used in EC to generate samples from the population. These mechanisms are based in the adequacy of the sampling methods previously explained. The stratified versions of the classical mechanisms are included and their impacts on genetic diversity are discussed. The chapter include bibliographic and journal references in order to deepen the knowledge in each issue. Some journal references about successful implementations will be included too.