快递服务便利店选址问题研究

发布时间：2018-05-21 06:11

本文选题：便利店 + 选址　；参考：《清华大学》2014年硕士论文

【摘要】：近十年来，随着网上购物的兴起，快递行业也得到了迅速的发展，但快递行业一直给大众留下低端服务业的印象，因为在城市里，我们通常见到的只是骑着电动车到处送快件的快递员。因此建立规范统一、带有明显企业标志的快递服务便利店，可以提高快递企业形象，增强人们对快递业的认同感，同时快递企业还可开拓附加服务来增加营收，如自寄自取、快件包装等。零售选址方法主要分为综合分析法和数学模型法。综合分析法即综合分析候选位置点经济、交通、需求量、建设费用等信息，，并应用层次分析法对候选位置点综合评价，最终选出符合公司各项目标的位置点；这种方法得出来的结果一般更加合理，但同时这种方法也非常耗费人力、物力，且费时较长。由于本文所研究选址问题需设立便利店数目较多，单个便利店设立成本相对较低，因此基于点需求和面需求，本文建立两个多目标选址模型来解决快递服务便利店选址问题。在基于点需求的多目标选址模型中，本文将客户需求点分为寄件业务需求点和取件业务需求点，针对两种业务不同特性在模型中做不同假设和处理，然后将市场份额与选址费用作为模型目标，并考虑门槛约束建立模型。利用某快递公司在上海徐汇区一周内收发快递数据作为模型的输入数据，利用多目标进化算法MNSGA-II和NSGA-II进行求解，分别得到一组帕累托非支配解集。根据本文采用的四个多目标优化评价指标对这两种算法进行比较，结果显示MNSGA-II求解的质量略优于NSGA-II,但NSGA-II和MNSGA-II得到的非支配解集均很好地逼近了帕累托最优解集。在基于面需求的多目标选址模型中，利用软件工具ArcGIS根据道路和水系将需求区域切分成多个子区域，并计算这些子区域的需求密度。以各子区域中心点为便利店候选位置点，考虑道路、水系的阻隔因素计算各候选点覆盖需求总量。考虑门槛约束和便利店之间距离限制，以市场份额和选址费用为目标建立模型。模型利用NSGA-II算法进行求解，在求解质量和求解速度方面均达到很好效果。在基于点需求的多目标选址模型和基于面需求的多目标选址模型得到帕累托非支配解集后，均采用一种基于逼近理想解排序的方法（TOPSIS）给出最合理解，以供决策者参考。另外，在实际选址决策时，决策者可根据自身条件在帕累托非支配解集中选择合适的选址方案。
[Abstract]:Over the past decade, with the rise of online shopping, express delivery industry has also been rapid development, but the express industry has been leaving the impression of low-end services to the public, because in the city, What we usually see is the courier who rides an electric car everywhere to deliver express mail. Therefore, the establishment of a standardized and unified express service convenience store with a clear enterprise logo can enhance the image of express delivery enterprises and enhance people's identity with the express delivery industry. At the same time, express delivery enterprises can also develop additional services to increase revenue, such as self-posting. Express package, etc. Retail location method is mainly divided into comprehensive analysis method and mathematical model method. The synthetic analysis method is to analyze the information of economy, transportation, demand, construction cost and so on of the candidate location point synthetically, and to evaluate the candidate position point synthetically by using the analytic hierarchy process (AHP), and finally to select the position point which accords with the company's each target. The results obtained by this method are generally more reasonable, but at the same time, the method also consumes manpower, material resources and takes a long time. Because the number of convenience stores needed to be set up and the cost of setting up a single convenience store is relatively low in this paper, two multi-objective location models are established to solve the location problem of express service convenience store based on point demand and surface demand. In the multi-objective location model based on point requirement, this paper divides the customer demand point into send business requirement point and pick up business requirement point, and make different hypotheses and processing in the model for different characteristics of two kinds of business. Then the market share and location cost are taken as the model objectives, and the threshold constraints are considered to establish the model. Using a delivery company to send and receive express data within one week in Xuhui district as input data of the model, a set of Pareto non-dominated solution sets is obtained by using multi-objective evolutionary algorithms MNSGA-II and NSGA-II to solve the problem. The results show that the quality of MNSGA-II is slightly better than that of NSGA-II.However, the set of non-dominated solutions obtained by NSGA-II and MNSGA-II are all close to the Pareto optimal solution set. In the multi-objective location model based on surface requirements, the requirement area is divided into several sub-regions by using the software tool ArcGIS, and the demand density of these sub-regions is calculated according to the road and water system. Taking the center of each sub-region as the candidate location of convenience store, the total coverage demand of each candidate point is calculated by considering the blocking factors of road and water system. Considering the threshold constraint and the distance between convenience stores, the model is established with the aim of market share and location cost. NSGA-II algorithm is used to solve the model, which achieves good results in both quality and speed. After the multi-objective location model based on the point requirement and the multi-objective location model based on the surface requirement get the Pareto non-dominant solution set, the most reasonable solution is given by using a method based on the ranking of approximate ideal solutions, which is for the reference of the decision makers. In addition, in the actual location decision, the decision maker can choose the appropriate location scheme in the Pareto non-dominant solution set according to his own conditions.
【学位授予单位】：清华大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F259.2

【参考文献】