Distributed task allocation method for scalable heterogeneous swarm

Through local information exchange and collaborative decision-making, distributed unmanned swarms are capable of accomplishing complex tasks with high efficiency and autonomy. It demonstrates considerable potential for diverse applications. Task allocation, however, remains a critical challenge. For heterogeneous multi-aircraft coalition formation, the main difficulties lie in high computational complexity, substantial communication overhead, and the trade-off between solution quality and efficiency. To address these issues, we propose a distributed task allocation strategy based on a hierarchical architecture. First, a hedonic game-based self-organizing clustering algorithm is developed. Through distributed interactions, it enables cluster self-organization and assigns heterogeneous aircraft to task clusters according to their requirements. Second, the Hungarian method is extended to solve the coalition formation problem within each cluster. An implicit consensus mechanism and a task node splitting mechanism are introduced, allowing effective matching of aircraft to tasks. In this way, the specific requirements of different tasks for each type of aircraft are satisfied. Moreover, a cost estimation method is designed by integrating the aircraft dynamics model. It achieves a tight coupling between task allocation and trajectory planning, which ensures the dynamic feasibility of the allocation results. Simulation results show that the hierarchical strategy significantly improves scalability while reducing communication overhead without compromising solution quality. Finally, a gliding aircraft case study validates the feasibility of the proposed algorithm with coupled range estimation.