Energy Efficient Resource Allocation and Computation Offloading in Millimeter-Wave based Fog Radio Access Networks

Abstract

As the sophisticated applications with latency constrained are difficult to operate on mobile devices with lower computing capability, fog-computing based radio access networks (F-RANs) have become a research hotspot as a revolutionary network architecture. It can offload a proportion of user's input tasks and provide scalable computation services at the fog-computing access points (F-APs), which reduce the task processing time and extend the battery life of local devices. In addition, the incorporation of millimeter-wave (mm-wave) band further improves the network performance with the uploading rate greatly increased. In this paper, we consider the multi-user uplink scenario, and formulate the problem of minimizing the total energy consumption of all users within the required latency. On the one hand, the optimization of user association joint with sub-channel allocation is studied to determine the connection status between users and edge F-AP nodes. It is modeled as a two-sided matching game representing the resource competition among users. As a result, the sub-optimal solution is obtained at lower complexity. On the other hand, the optimization of computation offloading for all users is also performed. Local devices can adjust CPU computing speed using dynamic voltage scaling (DVS) technology, and the offloading ratio is solved by convex programming. Both yield closed-form solutions. To this end, performance evaluation under multiple system parameters demonstrates the lower energy consumption and higher effectiveness of our proposed scheme in comparison with the baseline schemes.