ABSTRACT
Efficient post-disaster emergency network design can improve rescue efficiency and service quality. Considering demand uncertainty, the evolution of disasters and supply penalty cost, a dynamic emergency network planning model based on distributionally robust optimization was established by introducing probabilistic constraint based on Wasserstein ambiguity set to guarantee satisfying service quality. This model can be reformulated into a mixed integer programming model and solved by Gurobi efficiently and accurately. Numerical experiments based on the emergency distribution network in some areas of Shanghai during the epidemic were conducted. The results prove that the distributionally robust model has more advantages in demand satisfaction and computational efficiency. The influence of Wasserstein radius on objective value and probabilistic constraint's satisfaction rate are also successfully explored.
