为了解不同类型网约车乘客的出行特征,帮助网约车司机制定合理的寻客策略,以北京市出租车网络订单数据及快车网络订单数据为基础,采用K-均值(K-means) 聚类方法,以乘客出行时间、行程时间、上车区域用地性质及下车区域用地性质作为特征变量,对两种网约车的订单数据进行聚类分析,并分别将其划分为4种需求类型。对出租车网络订单及快车网络订单的4种需求类型进行对比分析,发现二者的乘客出行需求呈现出相似的特征。其中有两类需求受乘客出行时间影响较大,工作日早晚高峰的出行需求更为活跃,乘客上下车区域的用地性质集中于混合用地性质。另外两类需求受乘客出行时间影响较小,乘客上下车区域集中于混合用地与居住用地、商业服务设施用地、绿地及广场用地3 种用地类型之间。对订单数据进行统计分析发现,快车网络订单行程时间集中在10~20min,出租车网络订单的行程时间集中在10~45min。快车订单以短时出行为主,当乘客行程时间较长时选择出租车的概率更大。
Abstract
In order to understand travel characteristics of passengers using different on-line car-hailing service and help the drivers to make a reasonable customer-search strategy, based on the on-line order data of taxis and express cars, the two kinds of order data were clusterly analysed and divided into four demand types by using K-means clustering method. Passengers’departure time, travel time, land use characteristics of the pick-up and drop-off districts were taken as characteristic variables. By comparing and analyzing the four demand types of on-line taxis and express cars, it was found that the passengers’travel demand of the two types presented similar characteristics. Conclusively, two types of demand were greatly affected by the passengers’departure time and the travel demand was more active in the rush hours of morning and evening on working days. Its pick-up and drop-off districts were concentrated in the mixed districts. The other two types of demand were less affected by passengers’departure time. Its pick-up and drop-off districts were concentrated between the mixed districts and the residential districts, commercial service facility districts, green space and square land. Statistical analysis of the two kinds of order data shows that the travel time of express car on-line orders was mainly 10~20 minutes and that of taxi on-line orders was 10~45 minutes. Express car orders were subject to short-term travels and the probability of choosing a taxi was greater when passengers had a longer travel time.
关键词
网约车 /
快车 /
K-means聚类分析 /
需求特征 /
订单数据
Key words
on-line car-hailing /
express car /
K-means cluster /
demand characteristic /
order data
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 韩一童,王卫安. 基于出租车GPS数据的城市居民出行时空分布研究[J]. 测绘与空间地理信息,2018,41(2):87-89.
[2] 陈泽东,谯博文,张晶. 基于居民出行特征的北京城市功能区识别与空间交互研究[J]. 地球信息科学学报,2018,20(3):291-301.
[3] 刘萌,邬群勇. 基于出租车OD数据的居民活动强度时空特征研究[J]. 福州大学学报(自然科学版),2018,46(2):204-209.
[4] 童晓君,向南平,朱定局. 基于出租车GPS数据的城市居民出行行为分析[J]. 电脑与电信,2012(1):56-59.
[5] 栾丽娜. 基于GPS数据的城市出租车运营分析与数据挖掘[D]. 济南:山东大学,2015.
[6] 齐林. 基于GPS数据的出租车交通运行特性研究及应用[D]. 哈尔滨:哈尔滨工业大学,2013.
[7] 覃正桃,赵靖,王家儒,等. 基于GPS数据的出租车交通运行特征研究[J]. 中国水运,2017,17(6):70-72.
[8] 杨扬,姚恩建,潘龙,等. 基于GPS数据的出租车路径选择行为研究[J]. 交通运输系统工程与信息,2015,15(1):81-86.
[9] 司杨,关宏志. 计划行为理论下出租车驾驶员寻客行为研究[J]. 交通运输系统工程与信息, 2016,16(6):147-152,175.
[10] ZHANG S H, WANG Z Y. Correction: Inferring Passenger Denial Behavior of Taxi Drivers from Large- Scale Taxi Traces[J]. PLoS One, 2017, 12(2): 1-21.
[11] CHEN D D, ZHANG Y, GAO L P, et al. The Impact of Rainfall on the Temporal and Spatial Distribution of Taxi Passengers[J]. PLoS One, 2017, 12(9): 1-16.
[12] 袁亮,吴佩勋. 城市居民对网约车与出租车的选择意愿及影响因素研究——基于江苏省调查数据的Logistic分析[J]. 软科学,2018,32(4):120-123.
[13] 赵道致,杨洁. 考虑公平目标的网约车服务价格管制策略研究[J/OL]. 控制与决策,2018:1-9[2018-06-14].
[14] 张永安,伊茜卓玛. 各地网约车政策评价与比较分析[J]. 北京工业大学学报(社会科学版),2018,18(3):45-53.
[15] GUO Y G, YANG Y, GUAN Y M. Study on the Game Relationship Between Online-Taxi and Traditional Taxi Under the Taxi-Hailing Apps[C]// Proceedings of the 2017 International Conference in Communications, Signal Processing, and Systems. Harbin: CSPS, 2017: 1639-1649.
[16] WONG R C P, SZETO W Y, WONG S C. A Cell-Based Logit-Opportunity Taxi Customer-Search Model[J]. Transportation Research Part C: Emerging Technologies, 2014, 48: 84-96.
[17] 池娇,焦利民,董婷,等. 基于POI数据的城市功能区定量识别及其可视化[J]. 测绘地理信息,2016,41(2):68-73.
基金
国家自然科学基金项目(51338008);国家自然科学基金项目(51378036)