舰载机高效出动与回收是航母航空作战能力生成的重要基础。甲板与机库之间的跨层转运调度直接影响舰载机保障与出动效率，但受有限空间、密集停放、多类设备资源和多阶段作业流程共同约束，存在资源竞争强、时序耦合复杂和空间冲突突出等难题。针对上述问题，本文对舰载机跨层转运调度问题进行了研究。综合考虑牵引车和升降机独占、转运阶段衔接、系留/解系留、停机位阻塞和路径避碰等约束，建立了以机群最大完工时间最小为目标的调度模型；设计舰载机出动顺序编码，通过解码过程完成时序安排、设备分配和空间冲突检测。鉴于所建模型具有多资源耦合、约束类型多样和资源瓶颈动态变化等特点，提出分布估计算法-置信上界策略超启发式算法（Estimation of Distribution Algorithm-Upper Confidence Bound strategy，EDA-UCB），利用分布估计算法学习低层启发式规则序列，并通过置信上界策略自适应选择局部搜索算子。仿真结果表明，在8机和12机转运算例中，所提算法均能够获得优于固定算子、随机算子和直接排序优化方法的调度方案，且收敛过程更稳定。本文所建模型能够描述舰载机跨层转运过程中的多资源耦合关系，提出的超启发式算法可提高舰载机在复杂约束条件下的转运调度能力。

Efficient launch and recovery of carrier-based aircraft constitute a fundamental basis for generating the aviation combat capability of aircraft carriers. Cross-level transfer scheduling between the flight deck and hangar directly affects aircraft support and sortie efficiency. However, this process is constrained by limited space, dense parking layouts, multiple types of equipment resources, and multi-stage operational procedures, leading to strong resource competition, complex temporal coupling, and prominent spatial conflicts. To address these issues, this paper investigates the cross-level transfer scheduling problem of carrier-based aircraft. A scheduling model is established to minimize the maximum completion time of the aircraft fleet, comprehensively considering constraints such as exclusive occupation of tractors and elevators, transfer-stage precedence, lashing/unlashing operations, parking-position blockage, and path conflict avoidance. An aircraft sortie-sequence encoding scheme is designed, through which temporal scheduling, equipment allocation, and spatial conflict detection are completed in the decoding process. Considering that the proposed model is characterized by multi-resource coupling, diverse constraint types, and dynamically changing resource bottlenecks, an Estimation of Distribution Algorithm–Upper Confidence Bound hyper-heuristic algorithm (EDA-UCB1) is proposed. The estimation of distribution algorithm is used to learn sequences of low-level heuristic rules, while the upper confidence bound strategy is employed to adaptively select local search operators. Simulation results show that, in both 8-aircraft and 12-aircraft transfer cases, the proposed algorithm obtains scheduling schemes superior to those generated by fixed-operator, random-operator, and direct sequence optimization methods, while exhibiting a more stable convergence process. The proposed model can effectively describe the multi-resource coupling relationships in the cross-level transfer process of carrier-based aircraft, and the proposed hyper-heuristic algorithm can improve transfer scheduling capability under complex constraints.