远程转场硬式空中加油能够显著提升运输编队的航程与持续作业能力，研究其任务分配方法对提高运输效率与安全性具有重要意义。针对远程转场硬式空中加油任务分配中加油区域规划与加油机调度高度耦合、编队内各受油机的受油需求差异显著的问题，建立了考虑编队内受油优先级排序的加油任务需求模型。以最小化加油额外耗时、加油机调度成本以及加油区域风险系数为优化目标，建立了考虑航路约束的协同任务分配模型。通过引入基于混合冲突协同机制的位置更新策略、基于精英保留与动态拥挤距离的档案维护策略，提出了一种基于混合冲突协同机制的改进多目标灰狼算法的空中加油任务分配方法并进行仿真验证。仿真结果表明，在不同参数与场景下，本文所提方法均能获得优良的任务分配方案，证明其可行性；与现有的分阶段任务分配方法相比，本文的分配方法使得综合适应度值降低了11.0%，证明了本文所提方法的有效性；与其他算法相比，本文算法使得综合适应度值降低了10.9%，平均最优解出现时间降低了41%，体现出该算法良好的优化性能。

Long-range transfer boom aerial refueling can significantly enhance the range and sustained operational capability of transport aircraft formations, and research on task assignment methods is of great importance for improving transportation efficiency and operational safety. To address the issues of strong coupling between refueling zone planning and tanker scheduling, as well as the significant heterogeneity in refueling demands among receiver aircraft within a formation, a refueling task demand model incorporating priority ordering of receiver aircraft is developed. With the objectives of minimizing additional refueling time, tanker scheduling cost, and refueling zone risk coefficient, a task assignment model considering route constraints is developed. By introducing a position update strategy based on the hybrid conflict–cooperative mechanism and an archive maintenance strategy based on elite preservation and dynamic crowding distance, an improved multi-objective gray wolf optimization algorithm based on the hybrid conflict–cooperative mechanism is proposed for aerial refueling task assignment and validated through simulations. Simulation results demonstrate that, under different parameters and scenarios, the proposed method can obtain high-quality task assignment solutions, verifying its feasibility and effectiveness. Compared with the existing sequential task allocation method, the proposed method reduces the overall fitness value by 11.0%, demonstrating its effectiveness and superiority. Compared with other algorithms, the new algorithm reduces the overall fitness value by 10.9% and shortens the average time to obtain the optimal solution by 41%, indicating strong optimization performance.