Identification of critical combination of vulnerable links in transportation networks - a global optimisation approach

This paper presents a global optimisation framework for identifying the most critical combination of vulnerable links in a transportation network. The problem is formulated as a mixed-integer non-linear programme with equilibrium constraints, aiming to determine the combination of links whose deterioration would induce the most increase in total travel cost in the network. A global optimisation solution method applying a piecewise linearisation approach and range-reduction technique is developed to solve the model. From the numerical results, it is interesting and counterintuitive to note that the set of most vulnerable links when simultaneous multiple-link failure occurs is not simply the combination of the most vulnerable links with single-link failure, and the links in the critical combination of vulnerable links are not necessarily connected or even in the neighbourhood of each other. The numerical results also show that the ranking of vulnerable links will be significantly affected by certain input parameters.