Dispatch-sequencing optimization for drone delivery services: Models, properties and cutting planes

This paper investigates the dispatch-sequencing optimization problem for drone delivery services in the emerging low-altitude economy, aiming to minimize total service time by jointly optimizing order assignment and sequencing across multiple drone stations. The proposed framework is applicable to a wide range of on-demand delivery scenarios, such as food delivery, medical transport, parcel handling, and document distribution. We first formulate a static version of the problem as a Mixed-Integer Programming (MIP) model, analyze its structural properties, and derive a set of optimality, dominance, and lower-bound-based cutting planes. These are embedded within a tailored branch-and-cut algorithm that achieves substantial acceleration: over 6,000 × average speedup on small instances (maximum exceeding 19,000 × ) with optimality proven in seconds, and over 104 × average speedup on medium-scale instances with average optimality gaps below 3.5%. To assess performance at practical scales and generate managerial insights, we develop a simulation system and extend the static formulation to a dynamic variant that accounts for real-time system evolution. Extensive simulation experiments reveal that per-station drone fleet size and in-station processing time (including packing, weighing, and launching) are critical determinants of overall throughput and responsiveness. Our results offer valuable guidance for scaling drone fleet sizes, balancing fleet investment, workforce levels, and demand levels in diverse operational settings.