Transportation Centrality: Quantifying the Relative Importance of Nodes in Transportation Networks Based on Traffic Modeling

The topology of transportation networks such as road and rail networks determines the efficiency and effectiveness of the corresponding transport systems. Quantifying the relative importance of nodes of such networks is vital to understand their dynamics. Centrality metrics which are used in network science often make the assumption that only the shortest paths contribute to the importance of the nodes. In traffic scenarios however, while most traffic would preferentially go though paths of least cost, paths which are costlier are not omitted entirely. In this work, we introduce a new centrality metric, transportation centrality, which considers all paths that go through a node, and uses Logit functions and path lengths to compute the traffic which goes through each path, which in turn is used in centrality calculation. Therefore, this metric can be calculated based on topology alone, while it can also utilise traffic data if this is available. We demonstrate the utility of this new centrality metric by considering the suburban transportation networks of Seoul and Delhi. We also analyse the influence of the sensitivity parameter of the Logit function in the calculation of transportation centrality. We demonstrate that the introduced centrality metric is useful in understanding the relative importance of nodes in transportation networks, including networks for which no traffic data is available.