一种改进NSGA-Ⅱ算法的多目标优化研究与应用  

Research and Application of Multi-objective Optimization for An Improved NSGA-ⅡAlgorithm

作　　者：钟飞[1] 徐丁宜 ZHONG Fei;XU Dingyi(School of Mechanical Engineering,Hubei University of Technology,Wuhan Hubei 430000,China)

机构地区：[1]湖北工业大学机械工程学院,湖北武汉430000

出　　处：《机床与液压》2024年第22期133-142,共10页Machine Tool & Hydraulics

基　　金：湖北省重点研发计划项目(2023BAB088)。

摘　　要：针对某军工企业某维修生产线的多目标优化问题,建立以最大生产线平衡率和最小转运路径长度为目标的优化模型。根据该模型的特点,设计一种改进的NSGA-Ⅱ算法。基于已成熟的动态拥挤度研究,提出左右拥挤度的概念,在总拥挤度相同时,淘汰左右拥挤度更小的个体,进一步保证种群的多样性。同时,为了避免种群陷入局部最优,采用邻域搜索法去除每代种群中重复的个体,并引用独立于种群的精英保留策略,避免优秀个体丢失。最后,通过实例验证了该算法的有效性,并采用改进NSGA-Ⅱ算法将该实例中的生产线平衡率提升了约25.29%,转运路径长度减小了约36.86%。Aiming at the multi-objective optimization problem of a maintenance production line of a military industrial enterprise,an optimization model with the maximum production line balance rate and the minimum transfer path length as the goals was established.According to the characteristics of this model,an improved NSGA-Ⅱalgorithm was designed.Based on the mature dynamic crowding research,the concept of left and right crowding was proposed,and when the total congestion was the same,the individuals with less left and right crowding were eliminated,to further ensure the diversity of the population.At the same time,in order to avoid the population falling into local optimum,the neighborhood search method was used to remove duplicate individuals in each generation of the population,and the elite retention strategy independent of the population was used to avoid the loss of excellent individuals.Finally,the effectiveness of the algorithm was verified by example,and by using the improved NSGA-Ⅱalgorithm,the balance rate of the production line in the example was increased by about 25.29%,and the length of the transfer path was reduced by about 36.86%.

关 键 词：多目标优化 改进NSGA-Ⅱ算法 动态拥挤度 左右拥挤度 邻域搜索法 

分 类 号：F403.3[经济管理—产业经济]