An Improved Biogeography-based Optimization Algorithm

Biogeography optimization algorithm is a new evolutionary algorithm formed by migration model,because of its unique migration mechanism;it has a strong information-sharing capability and makes better use of information of the groups.However,the utilization of information of the traditional gene migration is low,and this pattern can only use the existing information of the groups,it does not have the ability to develop new information,and the traditional migration patterns only use the information of one single individual’s one character,although individual migration patterns can be use an individual information,it improves the efficiency of information groups to a certain extent,and have a certain ability to develop new information,but but this method also brings the disadvantage of decline about algorithm accuracy;for the above two points,the paper improved the algorithm as follows:Firstly,according to the individual migration mode information uses the defect is inefficient,reference differential evolution mutation operator,this mutation operator can use four individual information one time,it will greatly improve the efficiency of the use of the information,and one of the four individuals is the best in group.It is benefit for leading the algorithm to search the direction of the best answer and speed up the convergence rate.Further more,we take an adaptive way in the setting of the zoom factor value to further improve the performance of this new individual migration operator.Secondly,for the disadvantages precision decline of using only individual migration patterns,new migration pattern combine the new individual migration with the traditional gene migration patterns to form a multiple migration operator,each migration pattern produce a new individual separately,and choose the relatively better one as the offspring,using the multiple migration operator’s algorithm is called biogeography based optimization based on multiple migration(HLBBO),the experiment results show that the algorithm can effectively improve the convergence speed and has high precision and stability.Finally,based on the HLBBO algorithm,using the idea of decomposition this algorithm is extended to a multi-objective algorithm called multi-objective biogeography algorithm based on multiple migration(MHBBO).Experiments show that for solving multi-objective optimization problem,the distribution of the solution of the MHBBO obtained is more uniform,more close to the true Pareto front.