SOURCE Journal of Communications and Information Networks, Vol: 4 No: 4 pp: 24-31, Dec. 2019
Published Date: DEC 2019
Aerial platforms and edge servers have been recognized as two promising building blocks to improve the quality of service (QoS) in space-air-ground integrated vehicular networks (SAGIN). Communication intensive tasks can be offloaded toaerial platforms via broadcasting, while computation intensive tasks can be offloaded to ground edge servers. However, the key issues including how to allocate radio resources and how to determine the task offloading strategy for the two types of tasks, are yet to be solved. In this paper, the joint optimization of radio resource allocation and bidirectional offloading configuration is investigated. To deal with the non-convex nature of the original problem, we decouple it into a two-step optimization problem. In first step, we optimize the bidirectional offloading configuration in the case of the radio resource allocation is known in advance, which is proved to be a convex optimization problem. In second step, we optimize the radio resource allocation through brute-force search method. We use queuing theories to analyze the average delay of the two tasks with respect to the broadcasting capacity and task arrival rate. The offloading strategies with closed-form expressions of communication intensive tasks are proposed. We then propose a heuristic algorithm which is shown to perform better than interior point algorithm in simulations. The numerical results also demonstrate that the aerial platforms and edge servers can significantly reduce the average delay of the tasks under different network conditions.