Biogeography Optimization Algorithm To Improve The Research And Its Application

The biogeography-based optimizer (BBO) is an intelligent optimization algorithmsimulating species’ activities, and is proposed based on species’ survival, reproduction, declineand extinction in nature. It has become the hot research field of intelligent optimizationalgorithm with its special search mechanism and good optimization ability. Although it isrelatively new, the existing researches show its good ability to solve optimization problems. Itdeserves more attention for further improvement.This paper is based on the BBO’s basic principle, including algorithm improvement and itsapplication. The main work is summarized as follows.(1) The biogeography-based optimizer based on mean value migration and cauchy mutation(MCBBO) is proposed to solve single objective problems. Mean value migration operator andcauchy mutation operator are main operators of the MCBBO. With the mean value migrationoperator, individual’s migration mode is changed from original random transfer to the globaloptimal individual and random individual’s common guidance to realize a more accurate andeffective migration. Meanwhile, the usage of cauchy mutation can enhance the MCBBO’s abilityto escape from local optimizer and to achieve a better state.(2) The multi-objective biogeography-based optimization with mean value migrationoperator (MVBBO) is proposed to solve multi-objective problems. In the MVBBO, the maingroup conducts evolution process together with the introduced elite group, and a pruning strategybased on ε-dominance relationship is adopted to realize the elite group’s maintenance. Thismethod can not only prevent the Pareto front’s degradation, but also can effectively ensure theuniformity of the Pareto front distribution.(3) The group path planning subsystem is designed to generate real and natural paths on thebasis of the group animation design system. MCBBO is applied to generate group paths. Therequirements and restrictions of the specific scene are analyzed, such as environment boundarylimitation, obstacle detection, collision avoidance, group motion stagnation, etc. Finally fourcases of group path planning are realized, including group converging, following, evacuating andspreading.