Multi-Agent Cooperative Game Based Task Computing Mechanism for UAV-Assisted 6G NTN

Unmanned Aerial Vehicle (UAV)-Assisted becomes promising prospects for the development of 6G non-terrestrial networks (NTN) due to their low cost, low latency, flexible coverage, and support for emergency response. UAVs utilize close-range multi-point cooperative computing to provide computational, storage, and communication resources for 6G NTN. However, insufficient resources and variable business requirements can lead to node overload or task timeout. Therefore, selecting collaborative nodes and optimizing data transmission paths between non-terrestrial nodes has become a formidable challenge. Due to the capacity constraints, power limitations, time-varying communication links, and increasing terminal demands, existing algorithms face difficulties in meeting the requirements of cooperative computing for non-terrestrial nodes. To solve the issue, this paper proposes a multi-task cooperative computing mechanism that focuses on node selection and path optimization. Furthermore, node selection and path optimization algorithm based on multi-agent proximal policy optimization (NSPO-MAPPO) is designed to optimize the latency-aware multi-task cooperative computing objective, thereby improving computational task execution efficiency and service quality. Experimental comparisons with other algorithms demonstrate the effectiveness of our method in minimizing latency, enhancing the task load balancing degree, and ensuring stable and reliable cooperative computing services for non-terrestrial nodes.