CAJ | 학술논문

为提高多目标粒子群算法在解决复杂多目标优化问题中的整体性能，提出一种多策略融合的多目标粒子群算法。该算法采用均匀化与随机化相结合的方式初始化种群，在粒子速度更新中新增一扰动项，运用简化的 k-最近邻方法维持档案以及对档案个体赋予生存期属性并动态调整生存期值。实验结果表明，在 GD 和 SP 性能指标上，本文算法与另外5种对等算法在 ZDT 和 DTLZ 系列测试问题上进行对比，其表现出了总体显著性的性能优势。
위제고다목표입자군산법재해결복잡다목표우화문제중적정체성능，제출일충다책략융합적다목표입자군산법。해산법채용균균화여수궤화상결합적방식초시화충군，재입자속도경신중신증일우동항，운용간화적 k-최근린방법유지당안이급대당안개체부여생존기속성병동태조정생존기치。실험결과표명，재 GD 화 SP 성능지표상，본문산법여령외5충대등산법재 ZDT 화 DTLZ 계렬측시문제상진행대비，기표현출료총체현저성적성능우세。
In order to improve the overall performance of multi-objective particle swarm optimization algorithm (MOPSO) in solving complicated multi-objective optimization problems,a multi-objective particle swarm optimization algorithm integrating multiply strategies (MSMOPSO)was proposed in the paper.A new initialization approach of combining uniformization and random-ization was adopted in the MSMOPSO.Secondly,a disturbance item was added to the particle’s velocity updating formula.Thirdly, a simplified k-nearest neighbor approach was applied to preserve the diversity of external archive.Finally,every non-dominated par-ticle in the external archive was assigned the property of lifespan and the lifespan value would be adjusted dynamically during the run of the MSMOPSO.The experimental results illustrate that the proposed algorithm significantly outperforms the other five peer competitors in terms of GD,SP on ZDT and DTLZ test instances set.