Resilience-driven road network retrofit optimization subject to tropical cyclones induced roadside tree blowdown

This article focuses on decision making for retrofit investment of road networks in order to alleviate severe consequences of roadside tree blowdown during tropical cyclones. The consequences include both the physical damage associated with roadside trees and the functional degradation associated with road networks. A trilevel, two-stage, and multiobjective stochastic mathematical model was developed to dispatch limited resources to retrofit the roadside trees of a road network. In the model, a new metric was designed to evaluate the performance of a road network; resilience was considered from robustness and recovery efficiency of a road network.

The proposed model is at least a nondeterministic polynomial-time hardness (NP-hard) problem, which is unlikely to be solved by a polynomial time algorithm. Pareto-optimal solutions for this problem can be obtained by a proposed trilevel algorithm. The random forest method was employed to transform the trilevel algorithm into a single-level algorithm in order to decrease the computation burden. Roadside tree retrofit of a provincial highway network on Hainan Island, China was selected as a case area because it suffers severely from tropical cyclones every year, where there is an urgency to upgrade roadside trees against tropical cyclones. The authors found that roadside tree retrofit investment significantly alleviates the expected economic losses of roadside tree blowdown, at the same time that it promotes robustness and expected recovery efficiency of the road network.