一种增强型多目标烟花爆炸优化算法

发布时间：2018-06-21 01:58

本文选题：烟花爆炸优化 + 多目标优化算法　；参考：《电子学报》2017年10期

【摘要】：现实中多目标优化问题的多样化和复杂化要求发展新的多目标优化算法.在混合多目标进化算法设计思想和新型进化模型的启发下,提出一种增强型多目标烟花爆炸算法eMOFEOA,该算法利用均匀化与随机化相结合的方法生成均匀分布的初始种群,为算法后续搜索提供较好的起始点;对烟花爆炸半径采用精细化控制策略,即不同世代的种群具有不同的爆炸半径,而且同一种群内部因个体支配强度的差异而具有不同的爆炸半径,以节省计算资源;利用简化的k-最近邻方法维持外部档案的多样性.本文算法与另5种对等比较算法一同在12个基准多目标测试函数上进行性能比较,实验结果表明eMOFEOA算法在收敛性、多样性和稳定性上具有总体上显著的性能优势.
[Abstract]:The diversification and complexity of multi-objective optimization problems in reality require the development of new multi-objective optimization algorithms. An enhanced multi-objective fireworks explosion algorithm eMOFEOAA is proposed based on the design idea of hybrid multi-objective evolutionary algorithm and a new evolutionary model. The algorithm uses the combination of homogenization and randomization to generate uniform initial population. It provides a better starting point for the subsequent search of the algorithm, and adopts a fine control strategy for the explosion radius of fireworks, that is, the population of different generations has different explosion radius. Moreover, there are different explosion radii within the same population because of the difference of individual dominating intensity, so as to save computational resources, and to maintain the diversity of external files by using the simplified k- nearest neighbor method. The performance of the proposed algorithm is compared with the other five peer-to-peer comparison algorithms on 12 benchmark multiobjective test functions. The experimental results show that the eMOFEOA algorithm has significant performance advantages in terms of convergence, diversity and stability.
【作者单位】：华东交通大学软件学院;河北地质大学信息工程学院;
【基金】：国家自然科学基金(No.61763010,No.61165004,No.61402481,No.61663009) 江西省自然科学基金(No.20114BAB201025,No.20151BAB207022,No.20161BAB202064) 河北省青年拔尖人才支持计划(冀字[2013]17号) 河北省自然科学基金(No.F2015403046) 科学计算与智能信息处理广西高校重点实验室(No.GXSCIIP201604) 江西省教育厅科技项目(No.GJJ12307,No.GJJ14373,No.GJJ150539)
【分类号】：TP18

【相似文献】