Online user association and computation offloading for Fog-enabled D2D network

Fog computing has recently emerged as a promising solution to liberate mobile devices from increasingly intensive computation tasks. In this paper, we study fog-enabled device-to-device (D2D) collaborative network where mobile users can dynamically and benefically share their computational resources among each other. D2D fog computing is employed to achieve energy-efficient computation execution in the network. To this end, a power consumption minimization problem with task buffer stability constraints is formulated. Then an online D2D user association and computation offloading algorithm is developed based on Lyapunov optimization. Specially, the optimal CPU frequency scheduling, D2D association and power allocation are determined in each time slot. Further, a [O(1/V), O(V)] tradeoff between the power consumption and the execution delay is established. Simulation results are provided to validate the theoretical analysis of the proposed algorithm.