With the rapid development of the low‐altitude economy, unmanned aerial vehicles (UAVs) have exhibited enormous potential in emergency rescue and material delivery fields due to their flexibility; however, individual UAVs are often constrained by endurance issues. This paper focuses on rescue scenarios in complex terrains such as mountainous forests and considers service distances, while designing a UAV‐Vehicle collaborative delivery model. It constructs a dual algorithm based on clustering optimization and large‐scale adaptive neighborhood search (ALNS‐Kmeans++), enabling rapid optimization of the problem. Using a benchmark dataset as a sample, the study compares Genetic Algorithms (GA), Tabu Search (TS), and Ant Colony Optimization (ACO). In terms of solution results (total distance cost), ALNS outperforms GA, TS, and ACO by 27.2%, 23.2%, and 23.4%, respectively. The ALNS‐Kmeans++ algorithm integrating the clustering strategy shows better comprehensive practicability. It achieves suboptimal solution result (total distance cost) with the smallest fleet size (7.0% worse than ALNS), but its computational speed is 5.5% faster than that of the ALNS algorithm. In the validation within a rescue case study within tropical rainforests in Yunnan, this delivery model achieves efficient collaboration among three rescue vehicles and six UAVs. Emergency rescue can achieve full coverage and cost reduction through the strategy of dynamic relocation of fewer vehicles and collaborative UAVs. By the clustering optimization and dynamic adjustment of the ALNS‐Kmeans++, the efficiency of rescue tasks can be enhanced. It provides a replicable solution for the scenario‐based implementation of the emergency rescue system.
Xue et al. (Thu,) studied this question.