A link criticality index embedded in the convex combinations solution of user equilibrium traffic assignment

Abstract Identification of critical components in transportation networks is an essential part of designing robust and resilient systems. Topological criticality measures are based on graph theory and are applicable in multiple domains including communication and social networks. However, the non-linearity of link performance functions in transportation systems does not allow a perfect domain transfer of topological measures. Hence, transportation researchers take traffic flow characteristics into account while developing criticality measures. In such approaches, typically, a network performance measure is selected, then links are removed one-by-one, and traffic demand is reassigned to the updated network to calculate the impacts of each link failure. This consecutive link removal procedure requires multiple assignments which create a computational burden, especially for large networks. Overall objectives of this paper are (1) to compare and contrast selected criticality measures, and (2) to develop a new measure to identify critical components of transportation network, considering both traffic characteristics and network topology. For this purpose, the user equilibrium traffic assignment formulation is utilized, and the convex combinations solution algorithm is exploited for identification of link criticality ranking within a single traffic assignment. The developed measure is named Link Criticality Index (LCI). The LCI is compared with the existing measures in the literature through three numerical examples. Pros and cons of the LCI and selected measures are discussed in detail. The results indicate the proposed link criticality measure provides a balanced ranking with respect to connectivity/redundancy as well as the traffic conditions in the network.