Abstract:The collaborative target allocation of multiple drones is a combinatorial multi-model, multi-constrained optimization problem, whose solution space rises exponentially with the number of drones. When from the perspective of UAV(Unmanned Aerial Vehicle)collaborative target allocation to solve the problem of UAV formation switching, the standard Hungarian algorithm of a single drone flight path is too long, lead to individual drones power drops rapidly, so compared with other drones will land early, and will lead to other drones waiting time in the process of formation switching is too long, and then affect the overall formation flight time. Considering that the working environment of the UAV is in three D space and the cooperativity of formation switching time, this paper proposes an improved Hungarian algorithm, which can reduce the maximum movement distance of the single UAV as small as possible to prolong the flight time of the whole formation. By comparing various algorithms, the proposed algorithm has better results in this problem, and can well solve the formation switching problem of multiple drones.