【作者】 韩伟；

【导师】 於文雪； 胡芳槐；

【作者基本信息】 东南大学 ， 软件工程（专业学位）， 2020， 硕士

【摘要】 随着电商的发展和社会的进步,物质交流变得十分频繁,传统的物流调度受到了较大的考验。一个好的配送系统对于整个物流企业来说,不仅可以节约大量的人力物力资源,还能整合地方性资源、提高生产效率、缩短转运周期等。其中,集中配送车辆的优化问题是当前物流企业亟需解决的事项。本文针对物流配送中动态变化大的特性利用数学优化模型合理规划车辆的运输路线以达到实时调度配送车辆,这将对物流配送企业具有十分重要的实用价值。论文主要工作如下:首先,对物流运输行业的相关背景和实际生产的要求进行了较为详细的阐述,对集中配送电商车辆实时调度的需求进行了分析,基于现有的车辆货物资源设计了电商车辆实时调度系统的整体系统架构和各个功能模块。其次,对电商车辆的实时调度问题设计了优化策略。该策略将问题分解为区域划分和路径规划两个子问题。针对区域划分问题,给出了基于k-means算法的车辆区域划分方法;针对路径规划问题,设计了带有软时间窗约束且时间成本按权重参与成本计算的多目标决策的车辆路径规划函数模型,采用了改进的遗传算法以优化车辆的行驶路径。再次,对集中配送电商车辆实时调度系统进行了实现。设计了数据库并实现了实时调度的算法,集成了系统的各个功能模块。最后,进行了系统测试。围绕车辆调度的需求进行了仿真性实验和验证,实验表明该系统能达到车辆调度的动态调整优化,达到了减少运输车辆和降低运输成本的目的。更多还原

【Abstract】 With the development of the e-commerce and social advancement,material exchange becomes very frequent,and the traditional logistics scheduling has been greatly tested.A good distribution system can not only save a large number of manpower and material resources,but also integrate local resources,shorten the transfer period and improve production efficiency for the whole logistics enterprise.Among them,the optimization of vehicle distribution lines is the main problem that logistics enterprises need to solve first.According to the characteristics of great dynamic changes in logistics distribution,this thesis uses mathematical optimization model to reasonably plan the transportation route of vehicles to achieve real-time scheduling and distribution of vehicles,which will be of great practical value to logistics distribution enterprises.The main work of this thesis is as follows:Firstly,this thesis expounds the relevant background of logistics and transportation industry and the requirements of actual production in detail,analyzes the demand of real-time vehicle scheduling of centralized distribution e-commerce and designs overall system architecture and each functional module of the real-time scheduling system for e-commerce which is based on the existing vehicle and cargo resources.Secondly,this thesis designs a optimization policy for the real-time scheduling of e-commerce vehicles.The problem in the policy are divided into two sub-problems:region division and path planing;In order to solve the problem of region division,a vehicle region division method based on k-means algorithm is proposed.In order to solve the problem of path planning,a multi-objective decision-making function model of vehicle path planning with soft time window constraint and time cost participating in cost calculation by weight is designed,an improved genetic algorithm is used to optimize the driving path of vehicles.Thirdly,the vehicle real-time scheduling system by centralized is realized.The database is designed and the algorithm of real-time scheduling is accomplished.All functional modules of the system are integrated.At last,this thesis completes the system test.The simulation experiment and verification are carried out around the demand of vehicle scheduling.The experiment shows that the system can achieve the dynamic adjustment and optimization of vehicle scheduling and the purpose of reducing the number of transport vehicles and transport costs.更多还原