Component Importance for Multi-Commodity Networks: Application in the Swedish Railway

Preparedness and resilience planning for critical infrastructure networks requires evaluating the impact to the network when its components are disrupted. We extend the well-studied problem of component importance measures in single-commodity networks to multi-commodity networks by integrating a flow based multi-commodity optimization model with a multi-criteria decision analysis tool. A three-stage approach to assessing critical component importance with multicommodity impacts on network vulnerability of one-at-a-time component disruptions is proposed. We analyze commodity-specific impacts on network performance of the Swedish railway system to rank critical links with respect to the 20 different commodity types transported in the network. We conclude that the two proposed metrics to measure vulnerability, unmet demand and link usage over 90 %, give complementary information regarding the vulnerability of the network. It is further concluded that the approach supports exploration of how the lack of redundancy and capacity in the network can give rise to high levels of unmet demand or increased link usage for specific commodities that are transported in the network, although the overall robustness of the railway system towards single link failures is high.