Abstract:Delivery drones are sensitive to weather conditions, yet, little attention has been paid to this factor. In this study, we first quantitatively analyzed the danger if we do not take weather into account in a truck-drone delivery system. Then, the impacts from winds and rains are modeled as time-dependent drone travel time and multi-no-fly time windows. We further formulated a MILP model. A heuristic approach is proposed to solve instances (up to 200 customers) that are more realistic. Finally, our numerical study illustrated that our model could significantly reduce the delivery makespan impacted by the bad weather condition, compared to the original model in which weather impacts were not considered.

Parallel Drone Scheduling Traveling Salesman Problem with Weather Impacts