Energy-Aware Charging Station Capacity Planning for UAV Logistics Networks Under Reliability Reserve

Abstract

Uncrewed aerial vehicle (UAV) logistics is receiving increasing attention in low-altitude transportation systems, but its deployment requires appropriate planning of charging infrastructure. Unlike conventional facility planning problems, UAV charging-station capacity planning must jointly account for spatially distributed logistics demand, heterogeneous UAV energy consumption, service-radius constraints, investment budgets, and operational reliability. This study develops a mixed-integer linear program (MILP) for energy- and reliability-aware UAV charging infrastructure planning. The model determines charging-station opening decisions, charger installation levels, demand allocation to stations, and unmet charging demand. The formulation converts UAV charging sessions into energy-equivalent capacity loads, allowing higher-energy missions to consume more charger capacity than lower-energy missions. In addition, a reliability margin is introduced to reserve part of the installed capacity for demand fluctuations, queuing, charger downtime, and other operational uncertainties. To evaluate the model, representative UAV logistics scenarios are generated, including urban-core, county-level, delta-mixed, mountain, and medical logistics networks. Numerical experiments across representative synthetic scenarios indicate that planning difficulty is scenario-dependent: dense urban scenarios face the highest capacity pressure, whereas medical logistics scenarios require reliability-oriented service protection. Because the tested scenarios jointly vary spatial configuration and planning parameters, these differences are interpreted as composite scenario effects rather than isolated topology effects. The results further indicate budget-demand threshold effects, energy-stress impacts associated with changes in weather and electricity prices, and investment-energy-service trade-offs across different charging technologies. Computational results indicate that the MILP can solve the tested representative planning-scale instances under the adopted computational settings. This study provides an optimization approach for energy-aware UAV charging infrastructure planning and discusses planning insights under the tested scenarios.