危险货物运输车辆路径鲁棒优化研究

发布时间：2018-04-23 13:29

  本文选题：危险货物 + 车辆路径　； 参考：《兰州交通大学》2015年硕士论文

【摘要】：危险货物运输车辆路径优化是保障危险货物安全运输的基础环节之一。由于危险货物的危险性要求危险货物运输必须考虑运输风险这一因素,因此危险货物运输车辆路径优化问题是一个多目标优化问题,具有一定的复杂性。要在危险货物运输网络中找到一条或多条风险、时间或其他属性值最小的有效路径,需要知道运输网络中各路段上相应的数据信息。然而,这些重要的已知数据中,部分数据具有不确定性。由于基础数据的不确定,所得到的危险货物运输路径也不确定,有时候当路段上的某一属性值有细微变化时,所得到的危险货物运输路径有可能变得不可接受。因此,在不确定环境下,科学合理的设计危险货物运输车辆路径,综合考虑运输风险和运输时间等因素,找到具有较强鲁棒性的“稳健”的车辆路径至关重要。本文以危险货物运输车辆路径问题为研究对象,通过分析将该问题分为单配送中心危险货物单车配送路径问题、单配送中心危险货物多车配送路径问题、多配送中心危险货物单车配送路径问题、多配送中心危险货物多车配送路径问题等4个子问题,然后各自建立了鲁棒性可调的不确定环境下危险货物运输车辆路径多目标Bertsimas鲁棒优化模型。针对单配送中心危险货物单车配送路径多目标鲁棒模型,设计了一种采用庄家法构造非支配个体,采用聚集距离保持进化群体分布性的多目标单亲遗传算法进行求解;针对单配送中心危险货物多车配送路径多目标鲁棒模型,给出了一种改进的多目标遗传算法进行求解;针对多配送中心危险货物单车配送路径多目标鲁棒模型,设计了一种“两阶段法”对模型进行求解,先通过第一阶段的全局搜索聚类方法找到各配送中心所服务的客户需求点,然后由第二阶段的多目标单亲遗传算法对每个配送中心进行依次求解;针对多配送中心危险货物多车配送路径多目标鲁棒模型,设计了一种将多个配送中心和多个客户需求点综合考虑的混合多目标遗传算法进行求解。论文针对每一个子问题都给出了一个算例,通过所设计的算法对算例求解,结果表明本文设计的多目标遗传算法能够找到具有不同鲁棒性的危险货物运输车辆路径的Pareto解集,这对决策者找到一条或多条相对“稳健”的危险货物运输路径具有一定的参考价值。
[Abstract]:The route optimization of dangerous goods transportation vehicle is one of the basic links to ensure the safe transportation of dangerous goods. Because the dangerous goods must consider the transport risk, the route optimization problem of dangerous goods transportation vehicle is a multi-objective optimization problem, which has certain complexity. In order to find one or more effective paths with minimum risk, time or other attribute values in the transport network of dangerous goods, we need to know the corresponding data information on each section of the transport network. However, some of these important known data are uncertain. Because of the uncertainty of the basic data, the route of dangerous goods transportation is also uncertain. Sometimes, when there is a slight change in the value of a certain attribute on the road, the route of transport of dangerous goods obtained may become unacceptable. Therefore, under the uncertain environment, it is very important to design the vehicle route of dangerous goods transportation scientifically and reasonably, and to find the "robust" vehicle path with strong robustness considering the factors such as transportation risk and transportation time. This paper takes the vehicle routing problem of dangerous goods transportation as the research object, and divides the problem into single distribution center, single distribution center, single distribution center, multi-vehicle distribution route problem of dangerous goods. Four sub-problems, such as the multi-distribution center dangerous goods single cycle distribution path problem, the multi-distribution center dangerous goods multi-vehicle distribution path problem and so on four sub-problems, Then, the multi-objective Bertsimas robust optimization models of vehicle paths for dangerous goods transport in uncertain environments with adjustable robustness are established respectively. Aiming at the multi-objective robust model of single distribution center, a multi-objective single-parent genetic algorithm based on clustering distance to maintain the distribution of evolutionary population is proposed. For the multi-objective robust model of multi-vehicle distribution path of dangerous goods in single distribution center, an improved multi-objective genetic algorithm is presented, and the multi-objective robust model of single-vehicle distribution path of dangerous goods in multi-distribution center is proposed. A "two-stage method" is designed to solve the model. Firstly, the first stage global search clustering method is used to find the customer demand points served by each distribution center. Then the second stage multi-objective single-parent genetic algorithm is used to solve each distribution center in turn. For multi-distribution center, multi-objective robust model of dangerous goods multi-vehicle distribution path is proposed. A hybrid multi-objective genetic algorithm (MIGA) is designed to solve the problem, which takes multiple distribution centers and customer demand points into consideration. In this paper, an example is given for each sub-problem. The results show that the multi-objective genetic algorithm designed in this paper can find the Pareto solution set of vehicle paths with different robustness. It has certain reference value for decision makers to find one or more relatively "robust" transport routes of dangerous goods.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U492.336.3
，

本文编号：1792163