Resource Allocation for Energy-Efficient MEC in NOMA-Enabled Massive IoT Networks

Integrating mobile edge computing (MEC) into the Internet of Things (IoT) enables the IoT devices of limited computation capabilities and energy to offload their computation-intensive and delay-sensitive tasks to the network edge, thereby providing high quality of service to the devices. In this article, we apply non-orthogonal multiple access (NOMA) technique to enable massive connectivity and investigate how it can be exploited to achieve energy-efficient MEC in IoT networks. In order to maximize the energy efficiency for offloading, while simultaneously satisfying the maximum tolerable delay constraints of IoT devices, a joint radio and computation resource allocation problem is formulated, which takes both intra- and inter-cell interference into consideration. To tackle this intractable mixed integer non-convex problem, we first decouple it into separated radio and computation resource allocation problems. Then, the radio resource allocation problem is further decomposed into a subchannel allocation problem and a power allocation problem, which can be solved by matching and sequential convex programming algorithms, respectively. Based on the obtained radio resource allocation solution, the computation resource allocation problem can be solved by utilizing the Knapsack method. Numerical results validate our analysis and show that our proposed scheme can significantly improve the energy efficiency of NOMA-enabled MEC in IoT networks compared to the existing baselines.

Automated summary

In this article, the integration of mobile edge computing (MEC) into the Internet of Things (IoT) to provide high quality of service is explored through the application of non-orthogonal multiple access (NOMA) technique for energy-efficient MEC in IoT networks. A complicated resource allocation problem is formulated and decomposed into separate radio and computation resource allocation problems, which are further optimized using matching, sequential convex programming, and Knapsack methods. The proposed scheme significantly improves the energy efficiency of NOMA-enabled MEC in IoT networks compared to existing baselines, as validated by numerical results.