Distribution, Collaboration and Coevolution in Asynchronous Search

摘要

Spatial distribution of individuals in evolutionary search combined with agent-basedinteractions within a population formed of multiple societies can induce new powerful modelsfor complex optimization problems. The proposed search model relies on the distribution of in-dividuals in a spatial environment, the collaboration and coevolution of individuals able to actlike agents. Asynchronous search process is facilitated through a gradual propagation of geneticmaterial into the population. Recombination and mutation processes are guided by the popula-tion geometrical structure. The proposed model specifies three strategies for recombinationcorresponding to three subpopulations (societies of agents). Each individual (agent) in thepopulation has the goal of optimizing its fitness and is able to communicate and select a matefor recombination. Numerical results indicate the performance of the proposed distributed asyn-chronous search model.