Energy Efficient Distributed Learning in Integrated Fog-Cloud Computing Enabled IoT Networks

We investigate resource allocation scheme to reduce the energy consumption of distributed learning (DL) in the integrated fog-cloud computing enabled Internet of things (IoT) networks. In the envisioned system, IoT devices are connected with the cloud server (CS) via multiple fog access points (F-APs). We consider that local models are trained at the F-APs based on the collected data from the IoT devices and the F-APs collaborate with the CS for updating the model parameters. Our objective is to minimize the overall energy-consumption of F-APs subject to overall computation and communication time constraint. Towards this goal, we devise a joint optimization problem of scheduling of IoT devices with the F-APs, transmit power allocation, computation frequency allocation at the F-APs and decouple it into two subproblems. In the first subproblem, we optimize the IoT device scheduling and power allocation, while in the second subproblem, we optimize the computation frequency allocation. We develop a conflict graph based solution to iteratively solve the two subproblems. Numerical results reveal a considerable performance improvement of the proposed solution in terms of energy consumption minimization over the existing solutions.