Multi-Sortie UAV Inspection Route Planning in Three-Dimensional Container Yards with Static Obstacles Using a Hybrid SA–ALNS–2OPT Algorithm

Abstract

To address the challenges posed by a large number of spatially distributed inspection points, dense obstacles, and limited UAV endurance in container yard inspection, this study investigates a multi-sortie route planning problem for a single unmanned aerial vehicle (UAV) operating from a fixed docking station. A mathematical model is formulated to minimize total flight time subject to obstacle avoidance and safety clearance requirements, per-sortie duration limits, and exactly-once visitation of each inspection point. A hybrid simulated annealing–adaptive large neighborhood search–2-opt (SA–ALNS–2OPT) metaheuristic based on a sequence-first, split-second strategy is developed. ALNS optimizes the giant-tour visitation sequence, simulated annealing guides candidate solution acceptance, and 2-opt refines the local route structure. The optimized sequence is subsequently partitioned using a Split procedure to generate endurance-feasible multi-sortie solutions. Computational experiments in a simulated three-dimensional container-yard environment, including repeated trials, ablation studies, benchmark comparisons, and additional methodological validation, show that the proposed method produces high-quality inspection routes while satisfying safety and endurance constraints and maintaining stable solution quality. The proposed approach provides a methodological framework for large-scale UAV inspection planning in automated container yards.