Bisection-SCA optimisation for energy-efficient NOMA-based cooperative wireless-powered mobile edge computing networks
Year: 2026
Type: article
Abstract: In this paper, a non-orthogonal multiple access based cooperative wireless-powered mobile edge computing architecture is proposed for energy constrained subscriber devices in Internet of Things networks. The system includes an access point and three subscriber devices, co-located with an edge server, and run a harvest-then-offload protocol. A decode-and-forward cooperative relay with maximum ratio combining is incorporated to tackle the double-near-far effect. An energy minimisation problem is then formulated over the eight decision variables that are optimised jointly, along with the resulting non-convex problem, they are solved with the proposed bisection-successive convex approximation algorithm. The major benefit of the algorithm is that it reformulates a difficult non-convex problem into convex sub-problems, which guarantees a global optimum in a predictable, logarithmic number of steps. This is an attribute that is not possible with existing methods (successive-convex-approximation and learning-based approaches) because they only converge to local optima which depends on the initial point. The simulation results show that the energy of the AP is reduced by 80.0%, the optimality gap is 0.000% when compared to an independent fmincon multi-start simulation, the system energy efficiency is improved by 128.1%, and the sum rate is increased by 10.8% compared to the unoptimised baseline. The baseline operating point is found to have 1.55 times sum rate advantage over time division multiple access and 1.23 times decode-and-forward cooperation gain. Monte Carlo evaluation of the deterministic case over 500 Rayleigh fading realisations shows a feasibility rate of 76.2%, mean energy saving of 73.0% and convergence time of 7 iterations for all feasible realisations.
Source: Next ElectroEngineering
Language: English
Topic: IoT and Edge/Fog Computing
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7220831701