基于云计算的公交OD矩阵构建方法

Estimation of bus origin-destination matrix based on cloud computing

OD矩阵作为交通客流数据可视化的重要方法,是公交系统评价和优化的重要依据.提出了一种基于云计算的OD矩阵构建框架,在此框架下利用云计算优异的存取速度及计算性能,实现了对公交历史数据的筛选、预处理、变换、数据挖掘、解释评价等操作.此外,还提出一种综合了追踪乘客出行行为规律和估计站点热度的OD矩阵构建模型来实现对乘客下车站点较为精确的估计.算例分析表明,根据该模型计算得到的各站点下车人数结果能较好地拟合实际出行OD矩阵.

As a significant method for bus data visualization,OD matrix provides basic data for evaluating and optimizing bus system. Based on cloud computing,a model for estimating OD matrix was proposed to analyze the data after filtering,pre-processing,changing and mining. To generate an OD matrix with high accuracy,a model was also provided for estimating OD matrix by comparing and combining data via analyzing data of each holding IC card holder and generating gravities for each station. The testing results of real bus data show that the proposed model can estimate debarking data accurately to match real OD matrix well.