Vehicle routing problem with drones and variable service times for agricultural virus monitoring

Agricultural regions have long faced significant economic losses due to the widespread diseases, leading to decreased crop yields. Recently, many regions have adopted trucks and drones to monitor diseases. To help the managers effectively schedule these trucks and drones, this paper studies a Vehicle Routing Problem with Drones and Variable Service Times. This problem involves scheduling a fleet of trucks and drones to perform monitoring tasks, aiming to maximize the information profit collected from monitoring agricultural diseases by drones. The information profit is characterized as an exponential function of the service time, a decision variable to be optimized, in each region, leading to a Mixed-Integer Nonlinear Programming formulation. For small to medium-sized instances, a mathematical heuristic algorithm is proposed–Benders decomposition with acceleration strategies is integrated for drone routing, and a heuristic method is employed for truck routing. We also develop a specialized hybrid heuristic algorithm for large-scale instances involving an Adaptive Large Neighborhood Search. Extensive numerical experiments demonstrate the computational benefits of the acceleration strategies and the specialized hybrid heuristic algorithms, as well as the managerial advantages of considering variable service times for increasing the information profit from monitoring agricultural diseases.