基于行程时间预测的物流运输车辆路径优化研究

发布时间：2018-02-27 04:35

本文关键词： VRP模型两阶段算法行程时间预测交通流诱导模拟退火算法　出处：《大连海事大学》2015年硕士论文　论文类型：学位论文

【摘要】：在现如今经济高速发展的社会,物流已经成为了社会经济高速腾飞的重要基石。在我国第十二个五年规划中,将物流行业的发展提上了国家战略层面,这个政策说明了物流行业的春天来了。纵观物流行业发展至今,困扰着无数物流人的两个大的问题,也是物流费用产生的核心源头,即运输费用和仓储费用。根据国际上相关权威机构的调查数据显示,中国的物流行业的运输费用占据了物流的总费用的50%以上,占据国民生产总值的20%～300%。这个巨大的数字比例意味着每年都有一笔天文数字的资金在浪费掉,所以对运输费用的研究就非常的有必要。国内外学者对物流配送过程中的车辆路径问题都做出了非常大、非常多、非常广的研究和讨论,也得出了很多非常优秀的结果。但是,一般确定客户及配送中心之间的车辆行程时间,主要是通过获取客户及配送中心的地理位置坐标,利用坐标两点间直线距离公式和车辆平均行驶速度之商计算,并没有考虑到现实中复杂的交通情况,当得出的车辆路径方案遇到交通拥挤等复杂情况时,就不能很好的按照方案得到预期的结果。本文在详细研究带时间窗的多车型车辆路径问题的基础上,给出了一种能够求解出贴近现实交通情况的客户及配送中心之间车辆行程时间的预测方法,即通过交通流诱导技术中驾驶员行为特性、行程时间预测技术以及Dijkstra最短路算法,利用调查获取的平均交通流量数据,预测并求出客户及配送中心之间的车辆行程时间,为后续车辆路径方案优化提供车辆行程时间数据支持。在求解模型的算法设计上,借鉴两阶段算法理念,结合驾驶员行为特性、行程时间预测、Dijkstra最短路算法和模拟退火算法,求解出符合要求的配送车辆路径方案。本文具体内容包括：首先,介绍了论文题目的背景以及国内外研究现状和选题的意义和目的,阐述了本文的创新点。其次,对车辆路径问题以及交通流诱导理论进行了详细的阐述。最后,建立了带时间窗的多车型车辆路径问题模型,借鉴两阶段算法理念设计求解算法,并利用计算机仿真软件进行编程,通过案例应用得出符合要求的车辆路径方案。
[Abstract]:In today's society with rapid economic development, logistics has become an important cornerstone for the rapid development of social economy. In the 12th five-year plan of our country, the development of logistics industry has been promoted to the national strategic level. This policy shows that the spring of the logistics industry has come. Looking at the development of the logistics industry so far, there are two major problems puzzling countless logistics people, and it is also the core source of the production of logistics costs. That is, transportation costs and warehousing costs. According to the survey data of relevant international authorities, the transportation costs of China's logistics industry account for more than 50% of the total logistics costs. This huge proportion of GDP means that an astronomical amount of money is wasted every year. So the research on transportation cost is very necessary. Scholars at home and abroad have made a very large, many, very extensive research and discussion on the vehicle routing problem in the process of logistics distribution. However, they have also got a lot of excellent results. In general, the vehicle travel time between the customer and the distribution center is determined, mainly by obtaining the geographical coordinates of the customer and the distribution center, using the straight line distance formula between the two points of coordinates and the quotient of the average vehicle speed. It does not take into account the complex traffic situation in reality. When the vehicle routing scheme is met with complex situations such as traffic congestion, We can not get the expected results according to the plan. Based on the detailed study of the multi-model vehicle routing problem with time window, In this paper, a prediction method of vehicle travel time between customers and distribution centers, which is close to the actual traffic situation, is presented, which includes driver behavior characteristic, travel time prediction technology and Dijkstra shortest path algorithm in traffic flow guidance technology. Using the average traffic flow data obtained from the investigation, the vehicle travel time between the customer and the distribution center is predicted and calculated, which provides the vehicle travel time data support for the subsequent vehicle routing scheme optimization. Based on the two-stage algorithm, combined with driver behavior, travel time prediction Dijkstra shortest path algorithm and simulated annealing algorithm, the vehicle routing scheme that meets the requirements is solved. The specific contents of this paper are as follows: first, This paper introduces the background of the topic, the significance and purpose of the research at home and abroad, and expounds the innovation of this paper. Secondly, the vehicle routing problem and the theory of traffic flow guidance are elaborated in detail. The model of multi-vehicle vehicle routing problem with time window is established, and the solution algorithm is designed by using two-stage algorithm, and the vehicle routing scheme according to the requirements is obtained by using the computer simulation software for programming.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F259.2

【参考文献】