作者: John DeLaurentis , William E. Hart , Lauren Ferguson
DOI:
关键词: Mutation (genetic algorithm) 、 Convergence (routing) 、 Self adaptive 、 Evolutionary programming 、 Mathematics 、 Component (UML) 、 Simple (abstract algebra) 、 Algorithm 、 Evolutionary algorithm 、 Mathematical optimization
摘要: We consider the convergence properties of self-adaptive evolutionary algorithms (EAs). The search component these EAs implicitly adapts step lengths in response to their efficacy for generating improving points. analyze a (1, λ)-EA with simpler mutation updates than are commonly used Evolutionary Strategies or Programming methods. Although have been analyzed by several authors, our analysis provides first exact proof an EA. Our experimental and theoretical demonstrates that this EA robustly converges optimum symmetric, unimodal problem.