Abstract: ‍ ‍The future trend of unmanned aerial vehicle （UAV） swarm technology involves deploying a large number of low-cost UAVs to accomplish various complex tasks through collaborative sensing， information sharing， and a coordinated division of labor. These swarms possessed high intelligence and autonomy， and they gradually emerged as the future direction of UAV swarm technology. High-precision positioning technology played a crucial role in the maintenance of swarm stability， avoidance of collisions， and achievement of target guidance. Among these technologies， UAV swarms utilized IoT technology combined with advanced positioning algorithms to achieve precise positioning and coordination in the air. However， this also led to the emergence of complex joint UAV deployment and resource allocation problems （JUDRA）. Hence， this study addressed the problem of optimizing UAV swarm positioning by proposing a more adaptable TDOA+AOA joint positioning framework as well as weak communication constraints between UAV swarms that were more closely aligned with practical application scenarios. By simplifying the complex UAV swarm resource optimization and deployment problem into a non-convex， non-concave min-max optimization problem with constraints and then decomposing it into master-slave problems， we used an improved Gibbs sampling algorithm for the master problem and a particle filtering algorithm for the slave problem. The proposed method effectively dealt with the complex relationships between multiple variables and achieved optimization at different levels. To validate the effectiveness and practicality of the proposed method， we experimentally verified its effectiveness in positioning performance under different positioning frameworks and communication constraints between UAVs. Moreover， by considering different numbers of UAV swarms and target uncertainty radii， we further verified the robustness of the algorithm， thereby demonstrating its wide applicability and reliability in practical use.