Stochastic Drone Fleet Deployment and Planning Problem Considering Multiple-Type Delivery Service

Drone delivery has a great potential to change the traditional parcel delivery service in consideration of cost reduction, resource conservation, and environmental protection. This paper introduces a novel drone fleet deployment and planning problem with uncertain delivery demand, where the delivery routes are fixed and couriers work in collaboration with drones to deliver surplus parcels with a relatively higher labor cost. The problem involves the following two-stage decision process: (i) The first stage determines the drone fleet deployment (i.e., the numbers and types of drones) and the drone delivery service module (i.e., the time segment between two consecutive departures) on a tactical level, and (ii) the second stage decides the numbers of parcels delivered by drones and couriers on an operational level. The purpose is to minimize the total cost, including (i) drone deployment and operating cost and (ii) expected labor cost. For the problem, a two-stage stochastic programming formulation is proposed. A classic sample average approximation method is first applied. To achieve computational efficiency, a hybrid genetic algorithm is further developed. The computational results show the efficiency of the proposed approaches.