This research presents a two-layer optimization framework for collaborative task allocation and path planning of unmanned aerial vehicles (UAVs) to address the challenges of environmental monitoring on complex island terrains. First, a high-precision 3D terrain model is created by combining cubic Bézier curves with Gaussian filtering, providing a realistic environment for algorithm evaluation. Second, the upper layer employs the improved ant colony optimization–discrete artificial lemming algorithm (IACO-DALA) framework to achieve comprehensive coverage task allocation using the minimal number of UAVs, constrained by their maximum ranges. Finally, the lower layer develops a multi-strategy enhanced artificial lemming algorithm (MSEALA) for path planning by integrating gray wolf optimization (GWO) with the artificial lemming algorithm (ALA). This approach combines chaotic mapping initialization, golden ratio local search, and B-spline smoothing techniques to optimize multi-objective paths. Experiments demonstrate that on the traveling salesman problem (TSP) test sets (eli76, rand300) and in custom island scenarios, the IACO-DALA framework requires an average of 6.2 and 9.6 UAVs, respectively, compared to 8–30 UAVs for the baseline algorithm, while achieving complete task coverage. The MSEALA algorithm shows a 29.93% improvement in overall path planning efficiency and an 18.06% reduction in computation time, significantly outperforming algorithms such as ALA and the hybrid chaos game and gray wolf optimization (HCGO). This framework offers an effective solution for real-time island environmental monitoring and provides significant engineering value for island ecological protection and environmental governance in the marine engineering field. • Terrain modeling with Bézier curves and Gaussian filtering for multi-UAV simulation. • A two-layer framework coupling task allocation and path planning for island monitoring. • Improved IACO-DALA framework for minimum-UAV full coverage under range constraints. • MSEALA algorithm combining IALA and GWO to enhance multi-UAV path planning.
Duan et al. (Wed,) studied this question.