基于大数据的公交调度规则研究

发布时间：2018-11-15 17:45

【摘要】：提升公交的服务水平是实施公交优先的有效方式。随着智能公交的发展,在运营过程中,公交系统产生大量的数据,为公交规划和管理部门制定决策提供依据。基于公交IC卡数据和公交GPS数据,通过对数据的处理,及数据的分析和挖掘,可以获得有效的客流信息和车辆运行中的信息,并做出相应的预测,为公交调度提供决策支持。本文首先介绍了对公交数据源的分析,介绍智能交通背景下的公交智能系统产生数据的模式；公交数据的预处理流程,以IC卡数据和GPS数据为主,介绍了两种数据的不同的数据预处理步骤。本文研究了对公交数据的数据挖掘及分析。首先介绍了数据挖掘的基本概念和常用算法；对于IC卡的数据挖掘,本文分别研究了基于IC卡的客流时段划分,一些客流指标的统计分析,以及基于BP神经网络的客流预测；对于公交GPS的数据挖掘,本文研究了公交GPS的路段匹配,公交在运行过程中的运行特性的分析,以及基于BP神经网络的公交运行时间的预测。基于对公交数据的挖掘,本文研究了公交调度,包括静态调度及调度问题。对于静态调度,本文主要研究了时刻表的编制部分,即首先根据IC客流数据,划分客流时段,基于乘客等待成本最小、拥挤度最小、公交公司运营成本最小,建立公交时刻表的编制模型,并根据遗传算法进行求解；对于动态调度,本文分别研究了异常事件下的车辆调度形式,以及实时调度中的站点调度和站点间调度,即动态滞站调度和公交信号优先的动态调度,减少串车和大间隔的发生,提高公交的服务水平。根据以上研究内容,本文给出实例分析,分别以实例验证了以上的数据处理、数据挖掘、分析,以及公交调度的研究内容。
[Abstract]:Improving the service level of public transport is an effective way to implement bus priority. With the development of intelligent public transport, the public transport system produces a lot of data in the operation process, which provides the basis for the public transport planning and management department to make decisions. Based on bus IC card data and bus GPS data, through the data processing, data analysis and mining, we can obtain effective passenger flow information and vehicle operation information, and make the corresponding prediction, and provide decision support for bus scheduling. This paper first introduces the analysis of bus data sources, and introduces the data generation model of intelligent bus system in the context of intelligent transportation. Based on IC card data and GPS data, the different data preprocessing steps of two kinds of data are introduced. This paper studies the data mining and analysis of bus data. Firstly, the basic concepts and common algorithms of data mining are introduced. For IC card data mining, this paper studies the division of passenger flow time based on IC card, the statistical analysis of some passenger flow indexes, and the prediction of passenger flow based on BP neural network. For the data mining of bus GPS, this paper studies the section matching of bus GPS, the analysis of the running characteristics of public transport during operation, and the prediction of bus running time based on BP neural network. Based on the mining of bus data, this paper studies bus scheduling, including static scheduling and scheduling problems. For static scheduling, this paper mainly studies the compiling part of the timetable, that is, according to the IC passenger flow data, the passenger flow period is divided, based on the minimum passenger waiting cost, the minimum congestion, the minimum operating cost of the bus company. The model of bus timetable is established and solved by genetic algorithm. For dynamic scheduling, this paper studies the vehicle scheduling form under abnormal events, and the station scheduling and inter-site scheduling in real-time scheduling, that is, dynamic stop scheduling and bus signal priority dynamic scheduling. Reduce the occurrence of train strings and large intervals, improve the level of public transport services. According to the above research content, this paper gives the case analysis, and verifies the above data processing, data mining, analysis, and bus dispatch research content.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.17

【参考文献】