Theoretical Perspective of Convergence Complexity of Evolutionary Algorithms Adopting Optimal Mixing

摘要: The optimal mixing evolutionary algorithms (OMEAs) have recently drawn much attention for their robustness, small size of required population, and efficiency in terms number function evaluations (NFE). In this paper, the performances behaviors convergence OMEAs are studied by investigating mechanism (OM), variation operator OMEAs, under two scenarios---one-layer two-layer masks. For case one-layer masks, population is derived from viewpoint initial supply, while time analyzing progress sub-solution growth. NFE then asymptotically bounded with rational probability estimating performing evaluations. empirical results indicate that proportional to both degree cross competition one-layer-mask case. models also sizing decided supply when disjoint masks adopted, high selection pressure imposed OM makes composition sub-problems impact little on NFE, requirement increases reverse-growth probability.