Multiple Traveling Salesman Problem with a Drone Station: Using Multi-package Payload Compartments

In this paper, we investigate the Multiple Traveling Salesman Problem with One Drone Station (mTSP-ODS). More precisely, given a fleet of trucks located at the central depot, several drones at a drone station, and a set of customers, we are interested in serving all customers exactly once either by a truck or a drone in the shortest possible time. In the mTSP-ODS, drones are at a drone station, which can be activated by a truck's visit to hand over some parcels, which are then delivered by the drones. We analyze the impact of serving multiple customers in single drone sortie versus several sorties delivering a single parcel. We formulate both cases as mixed-integer linear programming (MILP) models. We solve the models by the standard MILP solver Gurobi and an effective heuristic that we introduce. We evaluate the models and the heuristic through computational experiments on benchmark instances. According to the numerical results, in addition to the efficiency of the heuristic, we observe that multi-package payload of drones can show benefits under some conditions.