MultiKulti Algorithm: Migrating the Most Different Genotypes in an Island Model

Migration policies in distributed evolutionary algorithms has not been an active research area until recently. However, in the same way as operators have an impact on performance, the choice of migrants is due to have an impact too. In this paper we propose a new policy (named multikulti) for choosing the individuals that are going to be sent to other nodes, based on multiculturality: the individual sent should be as different as possible to the receiving population. We have checked this policy on different discrete optimization problems, and found that, in average or in median, this policy outperforms classical ones like sending the best or a random individual.