Self-organized task offloading and resource allocation in cognitive air-ground collaborative edge computing networks

The integration of drone base stations (DBSs), mobile edge computing (MEC), and cognitive radio into the terrestrial network has the potential to address the ever-growing computing demands. Such networks are regarded as cognitive air-ground collaborative mobile computing (CAGC-MEC) networks. However, considering the large-scale and dynamic features of the Internet of Things (IoT) network, the traditional centralized method possesses prohibitive complexity, and the scalability may be compromised. This paper investigates the self-organized task offloading and resource allocation methods to maximize the quality of experience (QoE) of IoTs, which is evaluated by mean opinion score (MOS). We formulate a joint optimization problem of DBS deployment, resource allocation, task offloading, and channel selection schemes. Following the coverage then optimization methodology, a local Voronoi partition-based deployment method is proposed to improve the coverage rate. On this basis, we derive the closed-form expression of the resource allocation scheme. Subsequently, the task offloading and channel selection sub-problems are solved by Lagrange dual decomposition and best response methods, respectively. The proposed algorithm can be executed under the conditions that the positions of IoTs are previously unknown and without a central controller. Extensive simulations verify that the proposed algorithm is superior to other benchmark algorithms.