Selecting Appropriate Scale-free Behaviors of Individuals in Environments Where They Can be Eliminated

Abstract

We previously proposed two types of population-based search methods that rely on spatial scale-free behaviors of individuals or temporal scale-free behaviors of them. In the previous study, the temporal scale-free behaviors of individuals were shown to be suitable for specific problems in which indi- viduals around the global optimum can be eliminated probabilistically and the spatial scale-free behaviors of them were shown to be suitable for problems in which such elimination never occurs. That is, the two types of scale-free behaviors are complementary. So, in this paper, we first assume problems for which we cannot know if the probabilistic elimination of individuals occurs in advance, and then propose a new search method that selects appropriate types of scale-free behaviors for individuals during the search- ing. Simulation results show that the proposed method has good search performance on average for such problems.