Evaluating Road Network Resilience under Recurring and Non-Recurring Incidents

Resilience is one of the most representative indicators in the transportation system analysis. Some of the researchers have evaluated the resilience utilizing a performance metric. However, the same weight of long-term and short-term performance leads to inaccuracy in resilience evaluation, and is particularly producing a hysteresis problem when the resilience is calculated with long-term performance only. In addition, few studies compared the road network resilience under recurring and non-recurring incidents. In this paper, we propose a metric for evaluating resilience that considers both long-term and short-term performance to overcome the hysteresis problem. Meanwhile, a resilience assessment framework is proposed for evaluating road network resilience based on the road topology and dynamic performance. In addition, we compared the resilience between recurring and non-recurring incidents using the log data of North 5th Ring Road in Beijing, China. The experimental results show that (1) the sensitive analysis demonstrates that appropriate importance factor β and period m of short term can effectively evaluate section resilience and overcome the hysteresis problem; (2) from the perspective of the road network, the minimal resilience of non-recurring incidents is approximately 35.42% lower than that of recurring incidents; and (3) to increase the resilience of the road network, the management should focus on improving the performance of relief road sections for recurring incidents and diverting the vehicles about to enter the entrance ramps in advance for non-recurring incidents.