The scheduling problem with delivery and observation in collaboration after the disaster

After the disaster, incremental disaster points, uncertain information on emergency needs, and changing weather impede emergency deliveries and increase the risk to delivery personnel. This paper introduces the scheduling problem with a delivery vehicle and a high-performance observation unmanned aerial vehicle (UAV) in collaboration (SP-DVOUC) to address the difficulties encountered in emergency delivery. In the SP-DVOUC, the routes of the UAV and the delivery vehicle depend on the information updates on points, needs, and changing weather. Considering the information updates, we solve the SP-DVOUC by the rolling-horizon approach. We formulate a MIP model and propose two acceleration strategies specific to post-disaster dynamics, an insertion heuristic designed to rapidly restore solution feasibility during dynamic updates, and a large neighborhood search algorithm for the fast rolling-horizon approach. After extensive experiments based on the Yushu Earthquake in China, the performance of the SP-DVOUC and the fast rolling-horizon approach is verified. In particular, our algorithm outperforms three algorithms used in other studies to solve similar problems. In addition, some suggestions for urgent delivery and algorithm parameterization are given.