Link Criticality Analysis Based on Reliable Shortest Path in Network with Correlated Link Travel Times

Link criticality assessment in a transportation network is of significance to traffic operators in that it can provide guidance on road maintenance scheduling and reconstruction prioritizing under failure. Also, link criticality can help improve a network's resilience, providing more efficient services. This paper proposes a reliable-shortest-path-based analyzing procedure to study edge betweenness centrality. A Lagrangian-relaxation subgradient-projection (LR-SP) algorithm for finding the most reliable path in a network is described. More specifically, a covariance matrix in the network is first decomposed, then the problem is reformulated to a convex minimization program. With dualization, the problem is further simplified and solved using a subgradient projection method. Link criticality analysis is then conveyed followed by experiments on a real transportation network that validate the effectiveness of the proposed procedure. With validated link importance, we can then prioritize maintenance and restoration activities, thus improving the network's performance.