公路运输超载关键影响因素的识别与分析

发布时间：2018-03-20 03:35

本文选题：公路运输　切入点：影响因素　出处：《合肥工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着我国公路运输业的快速发展,公路超载问题日益突出,受到全社会的广泛关注。超载运输不仅破坏了公路网,加速地面损伤、引发交通事故还扰乱运输市场秩序,从而造成大量人员伤亡和财产损失。目前,国内外对于超载的研究主要集中于分析超载的成因和行政治理措施,而未从公路运输量专项调查数据的角度,研究影响公路运输超载的因素并对其产生的影响进行定量分析。本文基于2015年公路运输量专项调查采集的公路货、客运车辆的基本信息、能耗信息和运输信息探讨影响公路货、客运超载的影响因素,为查超、治超工作提供理论依据。首先,介绍公路运输量调查方案及基于公路运输车辆基本信息、能耗信息和运输信息,从超载分布情况、运输生产、运输量区域分布方面分析了安徽省公路运输特征;其次,针对传统的二元Logistic回归难以实现高阶交互、无法将连续变量合理离散化的问题,采用二元Logistic-CART决策树结合的方法对公路货运超载关键影响因素进行识别,并提出治理对策。最后,综合运用单因素显著性检验、多重共线性诊断、Logistic回归对公路客运超载关键影响因素进行识别,并提出治理对策。研究结果表明,货运车辆超载影响因素为经营类型、标记吨位、经营模式、调查旬货运量、标记吨位/调查旬货运量交互作用、调查旬平均运距/调查旬货运量交互作用、标记吨位/调查旬平均运距交互作用、标记吨位/调查旬货运量/调查旬平均运距交互作用,并且二元Logistic-CART决策树结合模型的AUC较传统二元Logistic提高0.05,二者差异有统计学意义;客运车辆的超载影响因素为标记客位、调查旬客运量。
[Abstract]:With the rapid development of highway transportation in China, the problem of highway overloading is becoming more and more prominent, and it has received extensive attention from the whole society. Overload transportation not only destroys the road network, accelerates the ground damage, causes traffic accidents, but also disturbs the order of transportation market. At present, domestic and foreign researches on overload mainly focus on the analysis of the causes of overload and administrative measures, but not from the perspective of the special survey data of road traffic volume. Based on the highway cargo, the basic information of passenger transport vehicle, energy consumption information and transportation information, this paper discusses the influence of highway cargo based on the special survey of highway traffic volume in 2015. The influencing factors of passenger overloading provide theoretical basis for the investigation and treatment of overloading. Firstly, the investigation scheme of highway traffic volume and the basic information of highway transport vehicles, energy consumption information and transportation information are introduced, and the distribution of overload and transportation production are introduced. The characteristics of highway transportation in Anhui Province are analyzed in terms of regional distribution of traffic volume. Secondly, the traditional binary Logistic regression is difficult to achieve high order interaction and can not reasonably discretize continuous variables. This paper uses the method of binary Logistic-CART decision tree to identify the key influencing factors of highway freight overloading, and puts forward some countermeasures. Finally, the single factor significance test is used synthetically. Logistic regression is used to identify the key influencing factors of highway passenger overloading, and the countermeasures are put forward. The results show that the influencing factors of overloading of freight vehicles are management type, marked tonnage, management mode, and the quantity of goods transported in ten days is investigated. The interaction of marked tonnage / survey ten days freight volume, the interaction of average ten days freight volume / survey freight volume, the interaction of mark tonnage / survey ten days average distance, the interaction of mark tonnage / survey ten days freight volume / survey ten days average freight volume, the interaction of ten days' average freight volume / survey average freight volume, the interaction of mark tonnage / survey ten days' freight volume / survey average distance, The AUC of the binary Logistic-CART decision tree model is 0.05 higher than that of the traditional binary Logistic model, and the difference between them is statistically significant.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F542

【相似文献】