ViewFulltext-福州大学学报（自然科学版.doc

一种基于出租车轨迹数据的优质客源评价模型及实证研究 邱端昇，邬群勇，刘萌，孙梅 福州大学地理空间信息技术国家地方联合工程研究中心，空间数据挖掘与信息共享教育部重点实验室，福建省空间信息工程研究中心，福建 福州 350002) 摘要：大数据环境为探究出租车客源时空分布提供了新的解决方案和思路。论文以出租车轨迹数据为研究对象，提出并构建了一种基于出租车轨迹数据的优质客源模型。模型将出租车载客段，相邻载客前后空载段的半段定义为一次完整的出租营运行为，引入出租车开销与行程速度关系，计算出租车动态消耗变化，以出租车单位时间内的盈利来量化出租客源。以福州市2015年5月份出租车轨迹数据为例，根据模型对轨迹数据进行预处理与密度计算，分析出租优质客源在时间和空间的分布规律。从时间规律上看，可以分为三段：1-5时，总量小，逐渐减少；6-19时，总量较大，逐渐增加；20-0时，总量大，数量激增。按空间分布规律，将福州优质客源划分为三类：1-5时，覆盖范围小，少量集中分布在城市CBD和娱乐场所；6-19时，覆盖范围大，集中在福州火车站，其他位置分布均匀但密度低；20-0时，覆盖范围大，大量优质客源集中分布火车站与城市CBD。 关键词：出租车；轨迹数据；优质客源模型；时空分布 中图分类号：U491 A model of high quality passenger based on taxi trajectory and case verification QIU Duansheng, WU Qunyong, LIU Meng, SUN Mei National?engineering?research?center?of?Geospatial?Information?Technology,?Key?Laboratory?of?Spatial?Data?Mining??Information?Sharing of MOE,?Spatial?Information?research?center?of?Fujian,?Fuzhou?University,?Fuzhou,?Fujian,?350002,?China) Abstract: New solution and ideas for studying the temporal and spatial distribution of taxi trajectory is presented by big data. Taking taxi trajectory as the research object, a new high efficiency passenger model based on taxi trajectory is proposed and constructed. In the model, a complete taxi trip consists of three parts, namely the loading state, the half of hunting state before loading and the half of waiting state after loading. The model adopts the profit of taxi in per unit time to quantify the trip. In order to calculate the dynamitic cost of taxi in different situation, the relationship between travel speed and the cost (mainly refers to fuel consumption) of taxi is introduced. Taking the taxi data of May 2015 in Fuzhou as an example, the pre-process and density calculation of the trajectory data were carried out according to the model. The spatio-temporal distribution of the high efficiency passengers is analyzed. The high efficiency passengers can be divided into three situations. The first one is from 1am to 5am and the passenger quantity is small and s