【作者】 吴祥国；

【导师】 杨晓光； 张汝华； 刘岱宗；

【作者基本信息】 山东大学 ， 道路与铁道工程， 2011， 硕士

【摘要】 随着我国城市社会经济以及物质文化生活水平的不断提升,快速城市化、机动化现象愈演愈烈；加之,城市人口不断增加,居民交通需求与日俱增,城市交通拥堵状况日益加剧。大力优先发展城市公共交通系统是我国城市缓解城市交通拥堵状况的根本出路和必然选择。公共交通基础数据资源是进行科学、有效的城市公共交通规划与管理运营决策的基础保障和关键。而传统基础数据采集采用人工调查的方式进行,不仅耗费大量的人力、物力、财力资源,还存在着调查数据样本量低、数据精度差、不能及时进行数据更新的固有缺陷。随着现代信息技术的发展,大量高科技硬件设施如公交刷卡收费系统、公交GPS系统等应用于城市公共交通系统之中,而所采集到的数据信息并未进行深度的挖掘与应用。本文基于公交IC卡和GPS数据进行城市居民公交出行OD矩阵的推导与应用,则实现了深度的公交基础数据挖掘与应用。首先,针对城市公交IC卡收费系统的国内外相关研究应用状况进行综述分析,分别从城市公交IC卡的推广应用、公交数据仓库构建与数据挖掘技术、公交出行OD矩阵反推、公交运营与规划应用4个方面展开。总结城市公共交通基础数据的采集技术,主要分为传统人工调查采集技术和基于现代信息技术的公交基础数据采集技术2类。城市公交数据仓库构建与数据挖掘技术、城市公交居民出行OD矩阵推导是本文的核心与关键。基于城市公交IC卡、GPS数据源,分别建立了基于公交IC卡数据的数据仓库结构体系、基于公交IC卡和GPS数据的数据仓库结构体系；并在此基础上,构建了包含数据仓库模块、数据挖掘工具模块、数据逻辑分析模块、用户控制模块4层结构的数据挖掘结构体系。城市公交居民出行OD矩阵推导主要包括公交站间OD矩阵推导、公交出行OD矩阵生成、公交出行OD矩阵校验3方面的内容。其中,公交站点OD矩阵推导是本章的核心研究内容,分别从上车站点推导、下车站点推导、公交换乘分析以及站间出行OD矩阵生成4个方面展开,并详细阐述了推导上车站点的基于公交IC卡数据和调度信息的聚类分析法、基于公交IC卡和GPS数据的时刻匹配分析法,推导下车站点的公交站点吸引权法、基于单个公交乘客的出行链分析法。最后,基于已挖掘的信息进行公交运营与规划的决策应用分析。分别从城市公交动态特性分析、公交服务质量分析、公交出行特征分析3个方面进行。更多还原

【Abstract】 Along with continuously upgrade of Chinese social, economic and living level, rapid urbanization and motorization phenomenon are more and more serious. Moreover, the uninterrupted increase of urban population and traffic demand make urban traffic congestion status more and more severe. Giving great priority to the development of urban public transportation system is the fundamental way and inevitable choice to relieve urban traffic congestion. Public transportation basic data resources are the basic guarantee and key to carry out scientific and effective urban public transportation planning and management decision-making. However, traditional basic data collection method is artificial investigation, which not only takes a lot of manpower, material and financial resources, but also exist several inherent defects such as a low sample survey data, poor data accuracy and no data updated in a timely manner. With the development of modern information technology, a large number of high-tech hardwares such as bus IC card payment system, bus GPS system being used in the public transportation system, but the collected data hasn’t been mined and applied deeply. Based on the public transportation IC card and GPS data to conduct urban public transportation OD matrix inference and application, the paper has carried out public transportation basic data mining and application more deeply.Firstly, based on the research and application status of urban public transportation IC card at home and abroad, the literature review has been divided into four parts including urban public transportation IC card application, urban public transportation data warehouse and data mining technology application, research on urban public transportation trip OD matrix inference, urban public transportation operation and planning application based on the data mining information. The urban public transport data collection technology has been divided into traditional manual survey collection technology and urban public transport data collection techonlogy based on modern information technology two parts. Urban public transportation data warehouse and data mining technology and urban public transportation trip OD matrix inference are core and key. Based on public transportation IC card and GPS data sources, data warehouse structure system based on based on bus IC card and data warehouse structure system based on based on bus IC card and GPS data have been set up. And on this basis, the data mining structure system has been built including data warehouse module, data mining tool module, data logic analysis module and user control module four layers. Urban public transportation trip OD matrix inference has been divided into three parts including urban public transportation station OD matrix inference, urban public transportation trip OD matrix generation and urban public transportation trip OD matrix calibration. Among them, public transportation station OD matrix inference is the core content of the chapter, including public transport boarding station inference, public transportation alighting station inference and public transport transfer analysis and public transportation station OD matrix generation. The boarding station inference method is respectively cluster analysis method based on public transportation IC card data and scheduling information, time matching method based on public transportation IC card data and GPS data. And the alighting station inference method is respectively public transportation station attracting weight method, trip chain analysis method based on a single bus passenger.Finally, the public transportation operation and planning decisions have been analyzed based on the mined information. There are three parts such as urban public transportation dynamic characteristics analysis, urban public transportation service quality analysis and urban public transportation trip characterstics analysis.更多还原