城市物流车辆配送路径优化研究

发布时间：2018-05-16 21:22

本文选题：车辆路径问题 + 城市配送　；参考：《重庆交通大学》2015年硕士论文

【摘要】：随着我国城镇化进程的加快,城市规模越来越大,国家对物流业健康发展政策的大力支持,使得城市物流成为近年发展的热点。城市物流系统是物流系统按区域划分出的一种类别,城市物流是有货物运输、存储、流通加工、配送等七种元素组成,其中配送是城市物流的核心因素之一,从系统上来说,城市物流配送网络系统是城市物流系统的核心子系统之一。一个现代化的城市物流系统必须具有现代化的城市物流配送网络系统,完善优化城市配送网络就是以城市整体为本,按照客户的要求,把货物安全送达相应的地点,在这个配送过程中,尽量使配送的车辆数、运营的里程数、对城市环境的污染等降到最少,这也是典型的车辆路径问题。由于城市内客户点数量和布局、交通运输、政策管理等特点,使得城市配送车辆路径问题具有更多的限制和约束,所以结合城市配送实际,选择客户时间、道路行驶速度等因素,研究带时间窗的时变条件下城市配送路径优化问题(TDVRPTW,Time-dependent Vehicle Routing Problem With Time Windows)。本文通过查阅大量文献资料,首先总结了车辆路径问题的概念及有关算法,详细描述了城市物流配送系统的内涵及功能,并对城市配送系统的特征及形成关键因素进行了分析研究;然后基于城市物流配送的特点及车辆路径问题基本条件的界定,分析建立了带时间窗的时变条件下城市配送路径问题相关数学模型;最后对问题算法进行了研究,把聚类思想融入到路径优化算法中,设计了基于模糊聚类-人工蜂群两阶段启发式算法:第一阶段对客户群进行模糊聚类分析,打破原有行政区划,将客户群划分不同类别,降低原有问题的规模;第二阶段对原有人工蜂群算法中的蜜源选择概率公式进行改进,添加与迭代次数和蜂蜜优良比率有关的扰动因子,解决算法迭代后期可能陷入局部最优的缺点。使用Solomon设计的Benchmark Problems中的标准测试数据库r101测试数据,采用matlab编程,在算法第一阶段进行聚类仿真,在算法第二阶段对划分的类别进行具体路线规划实验仿真,并与一般人工蜂群算法运算结果进行对比。最后以重庆市天友乳业股份有限公司自营专卖店配送为例,通过具体分析,对其具体配送路线进行了优化。本文根据城市配送特点,提出符合城市配送实际的TDVRPTW问题数学模型,设计了模糊聚类-人工蜂群问题求解算法,通过实例分析验证了模型的有效性及求解大规模问题时方法的可行性,有助于根据城市配送特征,制定更为合理城市配送方案,为优化城市物流配送路线提供决策依据。
[Abstract]:With the acceleration of the urbanization process in China, the urban scale is more and more large, the country's strong support for the healthy development policy of the logistics industry has made the city logistics become a hot spot in recent years. The urban logistics system is a category divided by the regional logistics system, and the urban logistics has seven elements, such as cargo transportation, storage, circulation processing, distribution and so on. Distribution is one of the core factors of urban logistics. From the system, the urban logistics distribution network system is one of the core subsystems of the urban logistics system. A modern urban logistics system must have a modern urban logistics distribution network system, and the perfect optimization of urban distribution network is based on the whole city. In accordance with the requirements of the customer, the safety of the goods is sent to the corresponding location. In this process, the number of vehicles, the number of mileages operating and the pollution of the urban environment are reduced to the minimum. This is also a typical vehicle routing problem. The city is equipped with the characteristics of the number and layout of the customer points, transportation, policy management and so on in the city. Vehicle routing problem has more restrictions and constraints, so combining the actual distribution of urban distribution, choosing customer time, road speed and other factors, the problem of urban distribution path optimization (TDVRPTW, Time-dependent Vehicle Routing Problem With Time Windows) under the time window of time windows is studied. Firstly, the concept of vehicle routing problem and related algorithms are summarized, the connotation and function of urban logistics distribution system are described in detail, and the characteristics and key factors of urban distribution system are analyzed and studied. Then, based on the characteristics of urban logistics distribution and the definition of the basic conditions of the vehicle routing problem, the analysis has been made with time. The mathematical model of urban distribution path problem under the time varying condition of the window; finally, the problem algorithm is studied. The clustering idea is integrated into the path optimization algorithm, and the two stage heuristic algorithm based on fuzzy clustering and artificial bee colony is designed. The first stage of the clustering analysis of the customer group is carried out to break the original administrative division, and the customer group will be broken. Divide the different categories to reduce the size of the original problem; the second stage improves the probability formula of the nectar source selection in the original artificial bee colony algorithm, adds the disturbance factors related to the number of iterations and the good ratio of honey, and solves the problem that the algorithm may fall into the local optimal point in the later period of the iteration. The Benchmark Problems designed by Solomon is used. The standard test database R101 test data, using MATLAB programming, in the first stage of the algorithm clustering simulation, in the second stage of the algorithm of the classification of the specific route planning experiment simulation, and compared with the general artificial bee colony algorithm results. Finally, Chongqing Tianyou dairy industry Limited by Share Ltd distribution store distribution. For example, the specific distribution route is optimized through specific analysis. According to the characteristics of urban distribution, this paper puts forward a mathematical model of TDVRPTW problem which is in line with the actual distribution of urban distribution, and designs a fuzzy clustering and artificial bee colony problem solving algorithm. The validity of the model and the feasibility of solving the large-scale problem are verified by the case analysis. It helps to formulate a more reasonable urban distribution plan according to the characteristics of urban distribution, so as to provide decision-making basis for optimizing urban logistics distribution routes.
【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U116.2

【参考文献】