面向最可盈利性的出租车出行路径分析模型与算法

发布时间：2018-03-04 22:31

本文选题：出租车出行路径　切入点：最可盈利路径　出处：《华东师范大学》2012年硕士论文　论文类型：学位论文

【摘要】：出租车是城市公共交通的重要组成,是以巴士、地铁、轻轨为主的大规模客运交通的合理补充。建立一个基于可靠信息的、科学合理的路径规划方法,对于实现出租车行业乃至城市整体的持续健康发展都具有重要意义。出租车的移动信息容易获取,其流动情况能够反映城市的道路交通动态,因此出租车轨迹数据常用作浮动车数据源以实现实时智能交通服务。本文采用上海市四大出租车服务公司2009年在1天24小时内采集的1万多台车的GPS轨迹数据,研究探索出租车最可盈利路径提取问题。文章首先对GPS轨迹数据进行了符合分析需要的预处理,包括坐标转换,OD矩阵提取,提出一种准确性和可操作性上表现良好的最小距离和多项式拟合相结合的地图匹配算法,探讨了路径规划的典型算法迪杰斯特拉(Dijkstra)算法,进一步对其改进以求解最短路径,为后续最可盈利路径的分析统计提供指标和计量数据。随后,论文从以下两方面展开建模和算法研究：第一,利用计算机模拟,重建当天出租车服务的时空动态,根据提取的出租车出行OD矩阵和与上海市交通网络GIS图层位置配准GPS轨迹行程进行语义属性转换和基于路网的出租车时空信息属性提取,并划分时段对其进行运营时空特性分析。在此基础上建立了基于位置的出租车空载巡游路线模型,用以分析如何提高出租车利用率,降低出租车运营成本。第二,在出租车运营经济计量方面,搜集实际的运营方式和运作价格表作为建模基础,并通过搭乘出租收集实际费用情况和路况检验校正模型参数,为盈利度建模整合影响因子,同时考虑行程路线长度和行程盈利因素,完成基于最可盈利原则的出租车路线实时预测模型,并从历史数据中抽取三条不同时长的行程,与模拟路线进行了理论上的对比分析。论文各章内容如下：第一章,分析研究背景和意义,并对国内外出租车尤其是本文的研究重点相关研究进展进行评述,提出本文的研究目标和研究内容,确定论文结构。第二章,介绍城市智能交通系统及其相关技术,包括车辆定位技术和地理信息系统等,并在此基础上探讨出租车运营系统的数据基础和理论基础。第三章,阐述模型建立的数据采集和预处理步骤,对研究的相关技术——大地坐标转换、GPS定位误差模型、道路投影方式、地图匹配等技术进行探讨,并进行优化试验。第四章,主要介绍模型的算法实现流程及其实例应用。首先进行基于路网的城市出租车运营时空特性分析,然后按照道路出行强度建立出租车空载巡游路线模型；之后分析出租车的运营特征,提出盈利核算模型,同时探讨Dijkstra算法的实现方法,完成基于最可盈利原则的出租车路线模型。第五章,对全文进行概括总结和展望,回顾论文的研究内容和研究过程,指出论文的创新和不足,提出有待继续研究的工作。
[Abstract]:Taxi is an important component of urban public transportation, and it is a reasonable supplement to large scale passenger traffic, which is mainly bus, subway and light rail. A scientific and reasonable route planning method based on reliable information is established. It is of great significance to realize the sustainable and healthy development of taxi industry and the city as a whole. The mobile information of taxi is easy to obtain, and its flow can reflect the dynamic of urban road traffic. Therefore, taxi track data is often used as a floating vehicle data source to achieve real-time intelligent transportation services. In this paper, the GPS track data of more than 10,000 vehicles collected in 2009 by four major taxi service companies in Shanghai within 24 hours a day are used. In this paper, the problem of extracting the most profitable path of taxi is studied. Firstly, the GPS trajectory data are preprocessed to meet the needs of analysis, including the OD matrix extraction of coordinate transformation. A map matching algorithm combining minimum distance and polynomial fitting with good accuracy and maneuverability is proposed. The typical algorithm of path planning, Dijkstraan Dijkstraan algorithm, is discussed and further improved to solve the shortest path. This paper provides indicators and measurement data for the following analysis and statistics of the most profitable path. Then, this paper starts the modeling and algorithm research from the following two aspects:. First, using computer simulation to reconstruct the space-time dynamics of taxi services on that day. According to the extracted OD matrix of taxi trip and the location registration of GPS path with GIS layer of Shanghai traffic network, semantic attribute transformation and spatio-temporal information attribute extraction of taxi based on road network are carried out. On the basis of this, a location-based taxi no-load cruise route model is established to analyze how to improve the taxi utilization rate and reduce the taxi operating cost. Second, in the aspect of econometrics of taxi operation, collect the actual operation mode and operating price list as the modeling basis, and collect the actual cost and road condition check model parameters through the taxi rental. In order to model the profit degree, integrate the influencing factors, consider the route length and the profit factor, complete the real-time forecast model of taxi route based on the most profitable principle, and extract three different length trips from the historical data. A theoretical comparison with the simulated route is carried out. The chapters of the thesis are as follows:. The first chapter analyzes the background and significance of the research, and reviews the research progress of domestic and foreign taxis, especially in this paper, puts forward the research objectives and contents of this paper, and determines the structure of the paper. The second chapter introduces the urban intelligent transportation system and its related technologies, including vehicle positioning technology and geographic information system, and discusses the data basis and theoretical basis of taxi operation system. In the third chapter, the data acquisition and preprocessing steps of model establishment are described. The GPS positioning error model of geodetic coordinate transformation, road projection mode, map matching and so on are discussed, and the optimization experiments are carried out. Chapter 4th mainly introduces the algorithm realization flow of the model and its application. Firstly, it analyzes the space-time characteristic of urban taxi operation based on road network, and then establishes the taxi no-load cruise route model according to the intensity of road trip. After analyzing the characteristics of taxi operation, a profit accounting model is put forward, and the implementation method of Dijkstra algorithm is discussed to complete the taxi route model based on the most profitable principle. Chapter 5th summarizes and prospects the full text, reviews the research content and process of the thesis, points out the innovation and deficiency of the thesis, and puts forward the work to be continued.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：U491;F572.88

【参考文献】