Estimation and improvement of transportation network robustness by exploiting communities

Abstract

Throughout the past years, researchers increasingly study the resilience of transportation systems through the lens of complex networks. This model simplification has helped to identify bottlenecks for all kinds of systems, e.g., subway, railway, and road networks. Nevertheless, for large networks, with ten thousand and more nodes, standard complex network-based robustness analysis methods do not scale up well. In this study, we propose to estimate and improve the robustness of transportation systems by exploiting the presence of communities in complex network representations. A community, by definition, is densely connected inside, but loosely connected to other components in the system. Accordingly, the community structure and the induced edges connecting communities can help to orchestrate a framework for better analysis and protection of our transportation systems. Experiments on twelve real-world transportation systems demonstrate the efficiency and scalability of our novel community-based framework.