Abstract:
Urban transportation networks are among the most important infrastructures whose performance has profound effects on different aspects of citizens’ lives; so, assessment the vulnerability of these networks is very important. The performance of urban transportation networks derived from the interaction between demand and supply. The sudden change in supply or demand makes perturbation in network performance. Decreasing the vulnerability of an urban transportation network include two fundamental concepts; averting failures and planning to minimize its consequences. This research focuses on the second concept. For this purpose, the critical nodes whose elimination leads to plunge of network performance have been identified.
Besides, introducing indices for identification of critical parts, the comparison of behavior of urban transportation networks against malicious attacks and random failures, assessment of network vulnerability based on the size of the giant component, efficiency, and adaptive capacity, evaluating the centrality measures distribution, and investigation whether urban transportation networks are scale-free or not have been assessed in this study.
The results reveal that although the networks are different from geographic point of view, their general behavior against failures are the same. But, introducing a unique index to identify critical parts in all networks is impossible. In Isfahan, in the failures affecting less than 3 percent of nodes, weighted degree capacity is the best measure to identify critical nodes. Moreover, streets with high capacity are critical streets whose elimination decreases the performance of the network. It is worthwhile mentioning that elimination of congested streets has no striking effect on the performance of the network.
Urban transportation networks are robust against random failures and vulnerable against malicious attacks; moreover, their betweenness distribution follows law-rule. Based on these two phenomena, we can say that urban transportation networks are scale-free.
