Resource allocation for offloading-efficiency maximization in clustered NOMA-enabled mobile edge computing networks?,??

In this paper, the problem of resource allocation for multi-carrier clustered non-orthogonal multiple-access (NOMA)-enabled mobile edge computing (MEC) networks is considered. In particular, the goal is to maximize the network sum-offloading-efficiency (SOE) via joint subcarrier assignment, power allocation, and computing resource allocation (J-SA-PA-CRA) for all user clusters, under the partial offloading mode, and subject to transmit power, delay and computing resource constraints. However, the formulated J-SA-PA-CRA problem is non-convex, and happens to be NP-hard, and thus is computationally-prohibitive. In turn, a low-complexity solution procedure is proposed, which decouples problem J-SA-PA-CRA into two subproblems: (1) J-PA-CRA per (user cluster, subcarrier) pair, and (2) one-to-one subcarrier assignment. The first subproblem is optimally solved via a successive convex approximation algorithm. As for the second subproblem, the polynomial-time complexity Kuhn?Munkres (KM) algorithm is proposed for optimal subcarrier assignment. Also, the stable marriage matching (SMM) algorithm is devised for a sub-optimal subcarrier assignment solution, but with less complexity than the KM algorithm. Extensive simulation results are provided to validate the proposed solution procedure, illustrating that the KM-based (SMM-based) solution procedure efficiently yields the optimal (a suboptimal) network SOE, in comparison to the J-SA-PA-CRA scheme (solved via a global optimization package), while being superior to their OMA counterpart schemes (e.g. FDMA and TDMA).

Automated summary

This paper addresses the resource allocation problem for multi-carrier clustered NOMA-enabled MEC networks, aiming to maximize network SOE through J-SA-PA-CRA. A low-complexity solution is proposed by decomposing the problem into two subproblems, which are optimally solved using a successive convex approximation algorithm and the KM algorithm, with the SMM algorithm proposed as a less complex alternative for sub-optimal subcarrier assignment. Simulation results confirm that the proposed solution efficiently achieves optimal (suboptimal) SOE compared to the global optimization package, outperforming OMA schemes like FDMA and TDMA.