Restoration Scheduling for Post-disaster Road Networks Based on Resilience Optimization

To accelerate the restoration of road network connectivity by effectively repairing damaged road segments in a post-disaster network, the restoration scheduling problem for post-disaster road networks aimed at resilience optimization was studied. First, the traffic demand satisfaction ratio was used to measure road network performance, based on which two road network resilience indicators, that is, network performance resilience and recovery rapidity, were established. Second, a restoration scheduling bi-level programming model based on resilience optimization was proposed, where the upper-level model, that is, a multi-objective mixed-integer programming model, was used to determine the road restoration selection and identify the restoration sequence. The lower-level model, that is, a day-to-day traffic assignment model, was employed to simulate the dynamic evolution of road network traffic flow during the restoration period. Then, the tabu search algorithm and the Frank-Wolfe algorithm were adopted to solve the upper-level and lower-level models, respectively, and the optimal solution was obtained by the iteration of the two algorithms. Finally, the effectiveness of the proposed model and algorithm was tested using a case study. The results show that under the given restoration budget constraint and work crew constraint, the optimal restoration scheduling generated by the proposed method can increase the road network resilience to the full extent. During the restoration process, only after all damaged road segments of a certain line in the affected area are restored, can the road network performance begin to increase in a stepped manner, and the speed of the road network performance improvement is low first and then high. Sensitivity analysis of different parameters shows that, with an increase in restoration budget, the performance resilience and recovery rapidity of the road network increase and decrease at average rates of 15.65% and 17.72%, respectively, and only increasing the restoration budget may not obtain a better restoration schedule. When the decision maker's preferences change, the performance resilience and recovery rapidity of the road network have opposite trends, and the average change rates of the two resilience metrics are 5.96% and 4.48%, respectively. Increasing the number of repair crews can improve the performance resilience and the recovery rapidity; however, the marginal benefit of increasing the number of repair crews drop gradually from 0.11 and 0.43 to 0.01 and 0.02, respectively. © 2022 Xi'an Highway University. All rights reserved.