Non-Orthogonal multiple access–enhanced mobile edge computing for the internet of things via task offloading and resource allocation

Abstract The exponential surge in computation-intensive applications within future fifth generation (5G) networks and industries has drawn considerable attention towards mobile edge computing (MEC). MEC has the capacity to allow wireless devices (WDs) to offload computational workloads to proximate edge servers (ESs). This paper scrutinizes the implementation of NonOrthogonal Multiple Access (NOMA)-enabled multi-access MEC, in which a WD can concurrently offload workloads to several ESs through NOMA, thereby reducing overall latency. To optimize this framework, we propose a joint optimization problem that allocates computation resources at ESs, distributes workloads, and allocates radio resources, all with the aim of minimizing task-completion latency and computational expenditure. Although the problem is inherently nonconvex, we capitalize on its layered structure to design an efficient algorithm that assures optimal solutions. Extensive numerical validation validates the efficacy of the proposed framework. Our approach significantly reduces delay and cost compared with traditional schemes while maintaining near-optimal performance with notably increased computational efficiency. We further extend the solution for multi-device environments, optimizing ES grouping to establish stable allocations that bolster system performance. These results confirm the superiority of NOMA-enabled multi-access offloading in enhancing MEC capabilities for internet of thing ecosystems.