基于支配度迁移模型的多目标生物地理学优化算法  

Multi-objective Biogeography Optimization Algorithm Based on Dominance Degree Migration Model

作　　者：黄浙铭 祁荣宾 HUANG Zhe-ming;QI Rong bin(Key Laboratory of Advanced Control and Optimization for Chemical Processes ,Ministry of Education,East China University of Science and Technology,Shanghai 200237, Chin)

机构地区：[1]华东理工大学化工过程先进控制和优化技术教育部重点实验室,上海200237

出　　处：《华东理工大学学报（自然科学版）》2018年第1期90-96,共7页Journal of East China University of Science and Technology

基　　金：国家重点研发计划项目(2016YFB0303401);国家自然科学基金青年项目(21506050);中央高校基本科研业务费;上海市自然科学基金(15ZR1408900)

摘　　要：针对传统多目标算法在解决MOPs问题时会出现Pareto前沿收敛结果不好、解集分布性不佳的情况,提出了基于支配度迁移模型的多目标生物地理学算法(MOBBO)。新的迁移模型充分利用了Pareto解之间的支配信息,有助于算法进行有效的个体评价和栖息地排序;为了强化算法的收敛效果,提出了基于优选特征库的自适应迁移策略,以便产生携带较好特征的候选解强化搜索能力;同时为了增强算法进化中Pareto解集的分布性,提出了改进的KNN密度估计方法淘汰过密的个体。通过ZDT和DTLZ系列测试函数以及MDI缩合过程的多目标问题优化上的比较,验证了MOBBO算法具有较快的收敛性和较好的分布延展性。Traditional multi objective evolutionary algorithm （MOEAs） may result in slow convergence on Pareto front and poor distribution of solution sets for MOPs. In order to deal with these shortcomings, this paper proposes a multi-objective biogeography optimization algorithm based on the migration model of dominance degree （MOBBO）. The proposed migration model makes full use of the information among the Pareto solutions so as to carry out the effective individual evaluation and the habitat sorting. In addition, this paper presents a self-adaptive migration strategy based on feature database for producing offspring with better features to strengthen the search ability. Meanwhile, in order to enhance the distribution of solutions during the evolution, this paper further modify the K-nearest neighbor （KNN） density estimation methods to discard overcrowded individuals. Numerical experiments on ZDT and DTLZ series and the condensation process of MDI verify the fast convergence and good distribution of MOBBO.

关 键 词：生物地理学优化 多目标优化 支配度迁移模型 自适应迁移 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]