为识别城市轨道交通网络关键站点并研究其多年演化，构建了基于截断奇异值分解(truncated singular value decomposition, TSVD)的关键站点识别方法，选取北京市2011—2019年早高峰时段的OD数据，通过关键特征向量分析网络客流演变并对城轨网络中关键站点进行识别，将其与复杂网络方法的识别结果进行对比。分析表明：TSVD法能很好地应用于考虑OD分布的网络关键站点识别，识别结果一定程度上更好代表了网络客流的空间分布。从识别结果看，北京轨道交通关键站点空间布局呈现多中心发展趋势，如西北西二旗，西南丰台科技园等站点逐步形成网络客流中心并相互联系；东南土桥、东北俸伯等站点也初步呈现网络客流中心的特征。

In order to identify key stations in urban rail transit networks and study their evolution over multiple years, we construct a key station recognition method based on truncated singular value decomposition (TSVD). OD data during the morning peak hours from 2011 to 2019 in Beijing are selected as the dataset. The analysis of network passenger flow evolution and identification of key stations in subway networks are conducted using the indicator of key eigenvectors and compared with the identification results obtained from complex network methods. The analysis demonstrates that, TSVD can be effectively applied to identify key network stations considering OD distribution, and the resulting identification outcomes to some extent better represent the spatial distribution of network passenger flow. In terms of application results, it is discovered that the spatial layout of key stations in Beijing's urban rail transit tends to develop towards multiple centers, such as the Northwest Xi'erqi, as well as the Southwest Fengtai Science Park. These stations gradually form network passenger flow centers and establish connections with each other. Stations like Southeast Tu Qiao and Northeast Fengbo also exhibit a preliminary trend towards becoming network passenger flow centers.

U231??????

北京交通大学大学生创新创业训练计划资助项目（2023100041703）