In the metropolises of China, the metro plays an increasingly important role in commuting because of its efficiency, affordability, and cleanliness. This paper attempts to explore the relationship between walking acce...In the metropolises of China, the metro plays an increasingly important role in commuting because of its efficiency, affordability, and cleanliness. This paper attempts to explore the relationship between walking access distance to metro stations and the demographic characteristics of passengers, such as age, monthly income, travel frequency, gender, and travel purpose, as well as the influence of the urban context. Nanjing Metro Line 2 is selected as the case study. By using different methods such as a questionnaire survey, spatial decay function, analysis of covariance(ANOVA), network analysis of routes, and K-means cluster analysis, it is suggested that demographic characteristics have a significant impact on the pedestrian walking distance, with the exception of gender. Furthermore, the paper finds a spatial decay effect in walking access distance, the decay rate of which, however, varies across stations. Terminal stations have a larger pedestrian catchment area than in regular and exchange stations. Moreover, the passengers of Nanjing Metro Line 2 can be classified into six groups according to their demographic characteristics, among which education and occupation are vital indicators in determining their willingness to walk to the stations. Middle-class passengers have a higher dependence on the metro and tend to walk longer than other groups do. This study provides an important reference for planners and transport sectors to optimize land-use and transport infrastructures.展开更多
针对城市轨道交通OD客流量短时预测问题,提出基于向量自回归(Vector Auto Regression,VAR)和动态模式分解(Dynamic Mode Decomposition,DMD)的VAR-DMD组合预测模型.首先,以北京市范围内的地铁站点为例,基于自动售检票系统数据(Auto Fare...针对城市轨道交通OD客流量短时预测问题,提出基于向量自回归(Vector Auto Regression,VAR)和动态模式分解(Dynamic Mode Decomposition,DMD)的VAR-DMD组合预测模型.首先,以北京市范围内的地铁站点为例,基于自动售检票系统数据(Auto Fare Collection,AFC),对地铁OD客流进行时空特征分析;其次,构建高阶加权向量自回归模型捕获OD客流数据的时空关联性,利用动态模式分解算法估算模型的参数,提取OD客流数据动态特征,实现数据的降维和降噪,利用实时更新算法更新模型的参数,实现长期连续预测;最后,以北京地铁AFC数据为算例,对模型进行验证.研究结果表明:相较于基准模型,VAR-DMD模型的运行时间减少96.67%,预测误差减少2.6%,具有较高的预测速度和预测精度,为城市轨道交通运营管理部门提供了可靠又及时的决策依据.展开更多
基金Under the auspices of National Nature Science Foundation of China(No.41701180)
文摘In the metropolises of China, the metro plays an increasingly important role in commuting because of its efficiency, affordability, and cleanliness. This paper attempts to explore the relationship between walking access distance to metro stations and the demographic characteristics of passengers, such as age, monthly income, travel frequency, gender, and travel purpose, as well as the influence of the urban context. Nanjing Metro Line 2 is selected as the case study. By using different methods such as a questionnaire survey, spatial decay function, analysis of covariance(ANOVA), network analysis of routes, and K-means cluster analysis, it is suggested that demographic characteristics have a significant impact on the pedestrian walking distance, with the exception of gender. Furthermore, the paper finds a spatial decay effect in walking access distance, the decay rate of which, however, varies across stations. Terminal stations have a larger pedestrian catchment area than in regular and exchange stations. Moreover, the passengers of Nanjing Metro Line 2 can be classified into six groups according to their demographic characteristics, among which education and occupation are vital indicators in determining their willingness to walk to the stations. Middle-class passengers have a higher dependence on the metro and tend to walk longer than other groups do. This study provides an important reference for planners and transport sectors to optimize land-use and transport infrastructures.
文摘针对城市轨道交通OD客流量短时预测问题,提出基于向量自回归(Vector Auto Regression,VAR)和动态模式分解(Dynamic Mode Decomposition,DMD)的VAR-DMD组合预测模型.首先,以北京市范围内的地铁站点为例,基于自动售检票系统数据(Auto Fare Collection,AFC),对地铁OD客流进行时空特征分析;其次,构建高阶加权向量自回归模型捕获OD客流数据的时空关联性,利用动态模式分解算法估算模型的参数,提取OD客流数据动态特征,实现数据的降维和降噪,利用实时更新算法更新模型的参数,实现长期连续预测;最后,以北京地铁AFC数据为算例,对模型进行验证.研究结果表明:相较于基准模型,VAR-DMD模型的运行时间减少96.67%,预测误差减少2.6%,具有较高的预测速度和预测精度,为城市轨道交通运营管理部门提供了可靠又及时的决策依据.
