Abstract
Offshore wind power substations serve as fixed bases for unmanned aerial vehicle logistics distribution. In the mixed-fleet scheduling of multiple unmanned aerial vehicle types, parking spot allocation, configuration decisions, and transfer timing are deeply coupled, rendering traditional scheduling algorithms ineffective for efficient solution. To address this, this paper takes the context of substations performing material delivery and inspection missions to surrounding wind turbines, maintenance vessels, and other offshore facilities. An optimization model is constructed that comprehensively considers group priority, aircraft type differentiation, and configuration transition factors, with objectives of minimizing total transfer time, number of configuration changes, and workload variance among transfer crews. An event-driven configurable greedy initial solution generation strategy is introduced, and an adaptive large neighborhood search algorithm is designed, incorporating domain-knowledge-driven destruction and repair operators, an adaptive weight mechanism, and a simulated annealing acceptance criterion. Experimental results demonstrate that the proposed algorithm achieves the optimal mean objective values across all six test cases, outperforming the conventional adaptive large neighborhood search algorithm by 7.0–7.9% and the genetic algorithm by 14.4–22.1%, while maintaining the lowest standard deviation. The algorithm exhibits robust stability and practical applicability, providing effective decision support for unmanned aerial vehicle logistics distribution scheduling at offshore wind power substations.