BBO Extensions
Haiping Ma, Dan Simon · 2017
This chapter discusses some extensions that can be made to biogeography-based optimization (BBO) to improve performance. It also discusses the effects of different migration curve shapes on BBO. The chapter provides some alternative approaches to BBO implementation and applies BBO extensions to solve some real-world optimization problems. In BBO, migration is a probabilistic operator that adjusts each solution xk by sharing features between solutions. If natural biogeography is an optimization process, then it stands to reason that modeling BBO more closely after natural biogeography could result in better optimization performance. In BBO, classical island biogeography theory corresponds to a monotonic decrease in immigration rate and a monotonic increase in emigration rate as solution fitness rank increases, although their shapes are different. The chapter proposes a migration operator called blended migration, which is a generalization of the basic BBO migration operator, and which is motivated by blended crossover in genetic algorithms (GAs).