User mode selection of NOMA based D2D communication for maximum sum-revenue

TL;DR

This paper proposes a dual-arm bandit model for cellular users to select between OMA and NOMA modes in D2D communication, maximizing revenue in 5G networks.

Contribution

It introduces a novel machine learning approach for mode selection in D2D communication, optimizing revenue through a dual-arm bandit model.

Findings

The proposed scheme effectively maximizes user revenue.

Simulation results confirm the model's efficiency.

Mode selection improves resource utilization.

Abstract

Device to Device (D2D) communication underlying cellular communication can improve the utilization efficiency of resources, and the cellular user equipment (C-UE) can utilize the rental resources for D2D user equipment (D-UE) to obtain revenue. In 5g communication, non-orthogonal multiple access (NOMA) is another promising technology. In this paper, cellular users can choose to apply orthogonal multiple access (OMA) mode or NOMA mode to rent their own resources to get the maximum benefit. We proposed a dual-arm bandit machine model to solve this mode selection game and proved the effectiveness of our scheme through simulation results.