联系我们

论文查询结果	返回搜索

论文编号：	11695
作者编号：	2120182899
上传时间：	2020/6/19 18:33:37
中文题目：	考虑客户满意度的区域配送中心选址-分配问题的多目标优化研究
英文题目：	Research on Multi-objective Optimization of Location-Allocation Problem of Regional Distribution Center Considering Customer Satisfaction
指导老师：	梁峰
中文关键字：	客户满意度；RDC选址；双层规划；多目标优化；启发式算法
英文关键字：	Customer satisfaction; RDC site selection; bi-level planning; Multi-objective optimization; heuristic algorithm
中文摘要：	近年来，我国大部分企业都面临经济形势下行和市场竞争加剧带来的发展瓶颈问题，为此物流作为企业的“第三利润源”被越来越多的企业所关注。区域配送中心（Regional Distribution Center ,RDC）建设是企业优化物流配送网络、降低运输成本的基础，因此本文以RDC建设选址和客户在配送中心间分配方案为基础研究问题，同时考虑到企业为提高竞争力，愈发注重物流配送服务带给客户的体验，因此在研究RDC选址-分配问题时考虑客户满意度和成本双目标。本文通过研究考虑客户满意度的双目标RDC选址-分配优化问题，为企业RDC选址决策提供参考。研究主要工作如下： 论文首先构建企业客户物流满意度计算模型和物流综合成本计算模型。本文将客户物流满意度概念引入企业RDC选址问题，通过计算客户对物流时间满意度和服务质量满意度，综合得出客户对企业的物流满意度。在计算物流综合成本时，充分考虑到面向企业级客户需要整车运输，因此运输配送成本与客户订货量呈阶梯函数性质。 针对RDC选址-分配问题特点，论文建立双层规划模型。上层模型分单纯和复合两种：单纯上层模型仅决策RDC选址，复合上层模型决策RDC选址、进货工厂和计划最大转运量；两种上层模型都以客户满意度和物流综合成本为双目标，因此求解结果为Pareto最优解集。下层模型决策客户在配送中心间的分配方法，以运输成本最小为目标。上层模型为下层模型提供RDC选址和最大转运量约束；下层模型为上层模型提供客户分配方案结果，帮助上层模型优化。 针对双层规划模型，本文分别优化设计二进制粒子群算法（BPSO）和带精英策略的快速非支配排序遗传算法（NSGA-II）进行求解。下层模型考虑大规模客户情况，使用BPSO求解，并使用按约束生成粒子方式使粒子保持在解空间、通过优化初始解提高算法效率；上层模型为单纯模型时采用遍历上层解空间方式求解，上层模型为复合模型时采用NSGA-II求解，并使用优生策略使个体保持在解空间，使用权重变化式分层选择操作提高局部搜索能力。 论文最后通过算例分析，求解出单纯上层模型时的RDC选址结果和复合上层模型时的RDC选址、进货工厂和计划转运量结果，证实模型和算法的有效性。
英文摘要：	In recent years, most enterprises in China have faced the development bottleneck problem brought about by the economic downturn and increased market competition. For this reason, logistics as a "third profit source" has attracted more and more enterprises' attention. The construction of Regional Distribution Center (RDC) is the basis for enterprises to optimize the logistics distribution network and reduce transportation costs. Therefore, this paper takes RDC construction site selection and customer allocation among distribution centers as basic research issues, while considering that enterprises are Competitiveness, more and more attention is paid to the experience of logistics distribution services to customers, so when researching RDC location-allocation issues, customer satisfaction and cost dual goals are considered. This paper provides a reference for enterprise RDC location decision-making by studying the dual-target RDC location-allocation optimization problem considering customer satisfaction. The main research work is as follows: The thesis first constructs the calculation model of enterprise customer logistics satisfaction and the calculation model of comprehensive logistics cost. This paper introduces the concept of customer logistics satisfaction into the RDC location problem of an enterprise. By calculating the customer's satisfaction with logistics time and service quality, the customer's logistics satisfaction is comprehensively obtained. When calculating the comprehensive cost of logistics, it is fully considered that the enterprise-level customers need to be transported by whole vehicles, so the cost of transportation and distribution and the order quantity of customers have a staircase function. Aiming at the characteristics of RDC location-allocation problem, the paper establishes a bi-level programming model. The upper-level model is divided into two types: simple and composite: the simple upper-level model only decides the RDC location, the composite upper-level model decides the RDC location, the incoming factory and the planned maximum transshipment volume; both upper-level models take customer satisfaction and comprehensive logistics costs as dual goals , So the solution result is Pareto optimal solution set. The lower model decides the customer's distribution method among the distribution centers, with the goal of minimum transportation cost. The upper layer model provides RDC location and maximum transshipment constraints for the lower layer model; the lower layer model provides the customer allocation plan results for the upper layer model and helps the upper layer model to optimize. For the bi-level programming model, this paper optimizes and designs the binary particle swarm optimization algorithm (BPSO) and the fast non-dominated sorting genetic algorithm (NSGA-II) with elite strategy to solve. The lower model uses BPSO to solve, and uses the method of generating particles according to constraints to keep the particles in the solution space and improve the efficiency of the algorithm by optimizing the initial solution; when the upper model is a simple model, it is solved by traversing the upper solution space, and when the upper model is a composite model, NSGA is used -II solve, and use the eugenics strategy to keep the individual in the solution space, and use the weighted variable layered selection operation to improve the local search ability. Finally, through the analysis of specific examples, the paper solves the RDC site selection result of the pure upper model and the RDC site selection, purchase factory and planned transshipment results of the composite upper model, which confirms the effectiveness of the model and algorithm.
查看全文：	预览  下载(下载需要进行登录)