Joint Traffic Admission, Resource Allocation and Mode Selection Protocol for NOMA-Based D2D Users Underlaying Cellular Network for 5G and Beyond Networks

Device-to-device (D2D) communications offload the data traffic from the cellular network and improve spectral and energy efficiency. Non-orthogonal multiple access (NOMA) is an emerging technique, which supports a large number of users through the same resource block (RB) with superposition coding and successive interference cancellation. In this paper, we jointly optimize the traffic admission, resource assignment, and mode selection scheme for NOMA-based D2D group (DG) underlaying a cellular network. In the considered DG, a D2D transmitter (DT) communicates with its respective D2D users (DUs) using NOMA protocol by reusing RB with cellular users (CUs). The problem of joint traffic admission, resource allocation, and mode selection is formulated as an optimization problem to maximize the time-average capacity of the NOMA-based DG while ensuring the time-average capacity of DUs and CUs within resource assignment constraints. To solve this problem, a distributed scheme based on Lyapunov optimization is designed to handle the admitted data traffic. Furthermore, to optimize the RB assignment and mode selection, we used the Greedy approach whereas, for power control, a successive convex approximation (SCA) technique is used. The numerical results demonstrate that the proposed algorithms improve the time-average capacity of the network in comparison to the conventional NOMA, and OMA-based D2D schemes.