Energy Aware Latency Minimization for Network Slicing Enabled Edge Computing

With the rapid growth of traffic in current wireless networks, the management of heterogeneous and time-critical services with limited computing and communication resources has become a significant challenge. Edge computing (EC) is one of the key techniques in enabling 5G networks to reduce the latency of services while the network slicing (NS) technique aims to provide users with tailored services by taking advantage of virtualization of a physical network. Besides, non-orthogonal multiple access (NOMA), one of the key enablers of 5G networks, enables more users to share wireless resources. We propose to leverage these techniques to minimize the total latency of the computing tasks with energy constraints. The combination of NS with NOMA for EC can reduce the unnecessary allocation of resources and consequently enhance both the energy and spectral efficiencies of the system. We first propose a binary scheme for offloading computing tasks and formulate an optimization problem to minimize the total latency of completing the tasks. We then transform the optimization problem to facilitate a near-optimal solution. We also introduce a heuristic algorithm to reduce the computational complexity of solving the optimization problem. Finally, extensive simulation results have demonstrated the effectiveness of our proposed algorithm.