Multi-period vehicle routing problem with time windows for drug distribution in the epidemic situation

This paper investigates a novel drug distribution system for the epidemic situation by linking two separate models. The improved SEIQR spread model and the multi-period vehicle routing optimization model are integrated to fit the epidemic environment. The epidemic spread model is used to capture virus transmission characteristics and drug demand fluctuations. Given this, we formally describe and model how to minimize the total travel time, the relative psychological cost, and the economic cost in the multi-period by considering realistic features including timevarying demand, priority distribution, and temperature control. Furthermore, we adopt an epsilon-global optimization method with the outer-approximation scheme for yielding global epsilon-optimal solutions in small instances and propose a hybrid tabu search algorithm (HTS) to solve large instances. In the HTS algorithm, the initial solution is constructed using an improved CW algorithm and then several neighborhood operators are developed to handle intra-route and interroute operations, destroy and repair operators are embedded in the tabu search framework to facilitate better solutions. Finally, extensive test experiments are conducted to demonstrate the performance of our proposed methods. An empirical case study of Chongqing city, China indicates the effectiveness of the periodic dynamic update mechanism and optimization algorithms. The results of the sensitivity analysis provide some management implications regarding drug distribution optimization in the event of a sudden epidemic situation.