Clustering Analysis of Metro Network Components Based on a Completely Directed Model

Research on the components' clustering of metro networks is a boon to urban transportation planning and development. The existing research on components' clustering cannot reveal the spatiotemporal features of directed platforms at the same station and directed tracks between two adjacent directed platforms. This paper first establishes a completely directed model to describe the metro system, where nodes and arcs represent directed platforms and tracks, respectively. The demand-supply level of directed platforms and tracks is evaluated by waiting time and in-vehicle crowdedness, respectively. After that, clustering analysis is conducted on the directed components to improve the metro’s operational efficiency and enhance the passengers' travel experience. Historical data on the Shenzhen subway system was used to validate the superiority of the proposed method, and the results showed that the demand-supply level of metro network components is dynamically changing. The clustering of platforms and tracks reveals the spatiotemporal characteristics of their respective categories, which also provides a basis for urban planning and development.