Trade-Off Analysis of NOMA-D2D and OFDMA-D2D Systems: Resource Allocation Perspective

Mutual interference between NOMA cellular users (CUs) and D2D links poses a challenge in extending NOMA’s resource allocation schemes to NOMA-D2D models. In this paper, we develop a novel resource allocation algorithm for NOMA-D2D systems and utilize it for comparing NOMA-D2D and OFDMA-D2D models. The proposed scheme aims to optimize the system sum rate and ensure a minimal rate for CUs. We formulate this as a mixed integer non-linear programming (MINLP) problem encompassing resource block (RB) allocation and power control. We propose a heuristic sub-optimal solution to lower the computational complexity. Our solution has two stages. In the first stage, D2D pairs and half of the CUs are admitted to the system to maximize the system sum rate. The rest of the CUs join the system in the next stage, while the upper bound for their achievable signal-to-interference-plus-noise-ratio (SINR) is maximized. Our first stage involves resource allocation for an OFDMA-D2D system. Employing an equal power allocation policy, we decompose the resource allocation problem for D2D pairs and CUs into two assignment problems. In the next step, the RB allocation scheme for CUs, which complies with the NOMA principle, is formulated as another assignment problem, and the power control strategy is updated. This two-stage heuristic solution allows a comparison of NOMA-D2D and OFDMA-D2D performances with identical power budgets, while the number of cellular connections in the NOMA-D2D system is twice that of the OFDMA-D2D system. Our results indicate that OFDMA-D2D achieves a higher sum rate and individual rates over NOMA-D2D.