基于时空矩阵的出租车OD需求可视化分析与SVD分解  

Visualization Analysis of OD Requirements and SVD Decomposition Based on Space and Temporal Matrix

作　　者：叶文博 YE Wenbo(Chang’an Dublin International Transportation School,Chang’an University,Xi’an Shaanxi 710064,China)

机构地区：[1]长安大学长安都柏林国际交通学院,陕西西安710064

出　　处：《交通节能与环保》2025年第2期80-85,共6页Transport Energy Conservation & Environmental Protection

摘　　要：本文以西安市出租车GPS数据为基础,深入分析了乘客的出行需求和出租车的时空分布规律。研究发现,每日早晚高峰出租车订单显著增加,晚高峰订单持续时间明显延长。轨迹点主要分布在市内主要道路上,周末需求量增加且向周边非主城区和县城扩展。通过奇异值分解算法对出租车GPS时空矩阵进行分解,得到时间分布矩阵与空间分布矩阵,这两个矩阵的关系揭示了总体需求趋势和不同时空下的出行需求差异。结果表明,模式1(常规需求模式)表现出白天需求高、夜间需求较低趋势,模式2(特殊需求模式)反映了下午和夜晚市区内的较高需求。Based on the GPS data of Xi’an taxi,this paper deeply analyzes the travel demand of passengers and the spatial-temporal distribution of taxis.The study found that the daily morning and evening peak taxi orders increased significantly,and the evening peak order duration significantly extended.The track points are mainly distributed on the main roads in the city,and the demand increases at weekends and expands to the surrounding non-main urban areas and county seats.The singular value decomposition algorithm was used to decompose the spatio-temporal matrix of taxi GPS,and the temporal distribution matrix and spatial distribution matrix were obtained.The relationship between the two matrices revealed the overall demand trend and the difference of travel demand under different spatio-temporal conditions.The results show that mode 1(conventional demand mode)shows the trend of high demand during the day and low demand at night,and mode 2(special demand mode)reflects the high demand in the afternoon and night.

关 键 词：运输规划与管理 出租车OD需求 SVD 数据可视化分析 GPS 时空分布 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]