基于密度聚类和NSGA-Ⅱ的无人车配送选址-分配优化  

Optimization of Unmanned Vehicle Delivery Site Allocation-allocation Based on Density Clustering and NSGA-Ⅱ

作　　者：朱志明 刘畅 袁凯 曹亚东[1,2] 孙知信[1,2] ZHU Zhi-ming;LIU Chang;YUAN Kai;CAO Ya-dong;SUN Zhi-xin(School of Modern Posts,Nanjing University of Posts and Telecommunications,Nanjing 210003;Post Industry Technology Research and Development Center of the State Posts Bureau(Internet ofThings Technology),Nanjing University of Posts and Telecommunications,Nanjing 210003;Anhui Yougu Express Intelligent Technology Co.Ltd.,Wuhu 241399,China)

机构地区：[1]南京邮电大学现代邮政学院,江苏南京210003 [2]南京邮电大学国家邮政局邮政行业技术研发中心(物联网技术),江苏南京210003 [3]安徽邮谷快递智能科技有限公司,安徽芜湖241399

出　　处：《物流工程与管理》2024年第11期70-74,40,共6页Logistics Engineering and Management

摘　　要：无人配送车被视为快递的“终极解决方案”,针对无人配送车的配送选址-分配问题,考虑综合成本和服务能力水平,建立多目标优化模型,通过密度聚类确定初始配送中心位置,采用结合密度聚类的NSGA-Ⅱ算法对需求点的分配进行优化,采用基于簇的种群初始化方法,引入Alpha Shape算法,加快收敛过程;采用簇边缘交叉和动态变异策略,使算法更具适应性,同时避免种群陷入局部最优解。以芜湖市为研究对象,收集菜鸟驿站POI数据进行实验,实验结果表明,改进后的算法能够较好地应用于无人配送车的配送中心选址-分配优化问题。Unmanned delivery vehicles are considered as the “ultimate solution” for express delivery.To address the distribution site-allocation problem for unmanned delivery vehicles,considering comprehensive costs and service capability levels,a multi-objective optimization model is established.Initial delivery center locations are determined through density clustering,and an NSGA-Ⅱ algorithm combined with density clustering is utilized to optimize the allocation of demand points.A cluster-based population initialization method is adopted,and the Alpha Shape algorithm is introduced to expedite the convergence process.Cluster edge crossover and dynamic mutation strategies are employed to enhance the adaptability of the algorithm and prevent the population from falling into local optima.Taking Wuhu City as the research object,POI data of Cainiao stations are collected for experimentation.The experimental results indicate that the improved algorithm can be effectively applied to the optimization problem of site allocation-allocation for unmanned delivery vehicles.

关 键 词：密度聚类 NSGA-Ⅱ 选址-分配优化 多目标优化 

分 类 号：F252[经济管理—国民经济]