【摘要】：目前,在城市交通領(lǐng)域,GPS設(shè)備已被廣泛應(yīng)用。例如,幾乎每輛出租車均安裝了GPS終端設(shè)備,這些設(shè)備會將出租車的行駛位置和運營狀態(tài)等數(shù)據(jù)傳輸給交通管理部門用于實時監(jiān)控。事實上,在出租車GPS數(shù)據(jù)中隱藏著出租車司機(jī)的群體智能信息如服務(wù)策略等,充分利用這些信息,一方面能夠引導(dǎo)出租車司機(jī)改善運營方式并增加收入,另一方面也能夠幫助管理部門提高出租車系統(tǒng)運營效率。然而,目前對出租車GPS數(shù)據(jù)的分析手段還相對初級;同時,出租車GPS數(shù)據(jù)規(guī)模龐大,例如,西安市一年的出租車GPS數(shù)據(jù)規(guī)模接近TB級。實踐證明,傳統(tǒng)的信息處理平臺已經(jīng)無法對上述規(guī)模的GPS數(shù)據(jù)進(jìn)行有效的分析。針對上述問題,本文基于大數(shù)據(jù)平臺Hadoop對出租車GPS數(shù)據(jù)進(jìn)行出租車服務(wù)策略的挖掘和分析。Hadoop是目前流行的一種大數(shù)據(jù)平臺,采用并行計算結(jié)構(gòu)能夠完成TB級以上規(guī)模的數(shù)據(jù)分析工作,同時由于開源、可擴(kuò)展、低成本、易于編程等優(yōu)點,已經(jīng)成為一個大數(shù)據(jù)處理領(lǐng)域的事實標(biāo)準(zhǔn)。出租車服務(wù)策略是針對司機(jī)服務(wù)過程建立的數(shù)學(xué)模型,主要指乘客搜索策略、乘客輸送策略、服務(wù)區(qū)域偏好等。在對GPS數(shù)據(jù)進(jìn)行數(shù)據(jù)預(yù)處理之后,首先提取出租車白班運營軌跡,并根據(jù)GPS軌跡對出租車司機(jī)的收入進(jìn)行量化;根據(jù)收入排名取出收入較高和一般的司機(jī)群體作為樣本司機(jī),提取并統(tǒng)計樣本司機(jī)的服務(wù)策略及使用情況,最后分析服務(wù)策略與收入的關(guān)系。在測試實例中,本文對樣本司機(jī)不同時隙在乘客搜索策略、乘客輸送策略、區(qū)域偏好上存在的異同點進(jìn)行了分析和比較。
[Abstract]:At present, in the field of urban transportation, GPS equipment has been widely used. For example, almost every taxi is equipped with GPS terminal equipment that transmits data such as taxi location and operation status to traffic management for real-time monitoring. In fact, the taxi drivers' intelligence information, such as service strategy, is hidden in the taxi GPS data. On the one hand, it can guide taxi drivers to improve their operation mode and increase their income. On the other hand, it can also help the management to improve the efficiency of taxi system operation. However, the current analysis of taxi GPS data is relatively preliminary; at the same time, taxi GPS data scale is huge, for example, Xi'an taxi GPS data scale is close to TB level. It has been proved that the traditional information processing platform can not effectively analyze the GPS data of the above scale. In order to solve the above problems, this paper uses big data platform Hadoop to mine and analyze taxi service strategy of GPS data. Hadoop is a popular big data platform. Using parallel computing structure can complete the data analysis work of TB scale and above. At the same time, because of the advantages of open source, expandability, low cost and easy programming, it has become a de facto standard in big data's processing field. Taxi service strategy is a mathematical model for driver service process, which mainly refers to passenger search strategy, passenger transportation strategy, service area preference and so on. After the data preprocessing of GPS data, the operation track of taxi day shift is extracted firstly, and the income of taxi driver is quantified according to the GPS track. According to the rank of income, the higher income and general group of drivers are taken as sample drivers, and the service strategy and usage of the sample drivers are collected and counted. Finally, the relationship between service strategy and income is analyzed. In the test example, the similarities and differences in passenger search strategy, passenger transport strategy and regional preference between different time slots of sample drivers are analyzed and compared.
【學(xué)位授予單位】：長安大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：U495;TP311.13

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 杜江;張錚;張杰鑫;邰銘;;MapReduce并行編程模型研究綜述[J];計算機(jī)科學(xué);2015年S1期

2 安實;匡偉明;;出租車GPS數(shù)據(jù)軌跡化方法研究[J];科學(xué)技術(shù)與工程;2015年11期

3 楊揚(yáng);姚恩建;潘龍;趙楠;;基于GPS數(shù)據(jù)的出租車路徑選擇行為研究[J];交通運輸系統(tǒng)工程與信息;2015年01期

4 劉智慧;張泉靈;;大數(shù)據(jù)技術(shù)研究綜述[J];浙江大學(xué)學(xué)報(工學(xué)版);2014年06期

5 陳吉榮;樂嘉錦;;基于Hadoop生態(tài)系統(tǒng)的大數(shù)據(jù)解決方案綜述[J];計算機(jī)工程與科學(xué);2013年10期

6 賈歐陽;阮樹驊;田興;楊峻興;李丹;;MapReduce中Combine優(yōu)化機(jī)制的利用[J];計算機(jī)時代;2013年09期

7 許丞;劉洪;譚良;;Hadoop云平臺的一種新的任務(wù)調(diào)度和監(jiān)控機(jī)制[J];計算機(jī)科學(xué);2013年01期

8 孟小峰;慈祥;;大數(shù)據(jù)管理:概念、技術(shù)與挑戰(zhàn)[J];計算機(jī)研究與發(fā)展;2013年01期

9 桂智明;向宇;李玉鑒;;基于出租車軌跡的并行城市熱點區(qū)域發(fā)現(xiàn)[J];華中科技大學(xué)學(xué)報(自然科學(xué)版);2012年S1期

10 姚遠(yuǎn);王麗芳;蔣澤軍;;HDFS一致性管理的研究[J];現(xiàn)代電子技術(shù);2012年08期

相關(guān)碩士學(xué)位論文 前5條

1 王鄭委;基于大數(shù)據(jù)Hadoop平臺的出租車載客熱點區(qū)域挖掘研究[D];北京交通大學(xué);2016年

2 周小玉;HDFS分布式文件系統(tǒng)存儲策略研究[D];電子科技大學(xué);2015年

3 趙利剛;基于出租車軌跡數(shù)據(jù)的載客情況可視化分析[D];浙江工業(yè)大學(xué);2014年

4 王謙;HADOOP作業(yè)啟動性能優(yōu)化實踐[D];北京交通大學(xué);2012年

5 童明;基于HDFS的分布式存儲研究與應(yīng)用[D];華中科技大學(xué);2012年

，

本文編號：2419782