Minimizing the total travel distance for the locker-based drone delivery: A branch-and-cut-based method

This article investigates the problem of minimizing the total travel distance in locker-based drone delivery, where the roofs of lockers are reused as parking platforms for drones. It is a drone routing and parking hybrid problem with modeling challenges. We find the sufficient and necessary conditions for feasible solutions and transform the original problem into a scale-tractable one. Subsequently, we propose a compact lower-bound formulation for the transformed problem and prove the total unimodality of the coefficient matrix. Furthermore, we develop a two-stage method in which a branch and cut algorithm solves the transformed problem and a heuristic constructs practical schedules for the original problem. Simulated tests demonstrate that the method can solve each simulated instance within one second. Random tests reveal that the method can efficiently solve instances with 1000 sites and 1500 tasks within an acceptable CPU time. A sensitivity analysis indicates that the complexity arising from routing flexibility is greater than that arising from parking flexibility.