QoE-driven resource allocation for D2D underlaying NOMA cellular networks

Device-to-device (D2D) communication can significantly improve network coverage and spectral efficiency. Meanwhile, non-orthogonal multiple access (NOMA) has recently been integrated with D2D communication to further improve connection density and satisfy explosive data rate requirements of end users. Considering quality of experience (QoE) has become an important indicator from the user perspective, in this paper, we study the QoE-driven resource allocation problem in a device-to-device (D2D) underlaying NOMA cellular network coexisting with D2D pairs and NOMA-based cellular users (CUs). Our target is to maximize the sum mean opinion scores (MOSs) of all users while guaranteeing the minimum QoE requirement of each CU and D2D pair, by jointly optimizing subchannel assignment and power allocation at CUs and D2D pairs. Since this problem is mixed-integer and non-convex, we first transform it into an equivalent yet more tractable form. Then, a two-stage iterative algorithm based on the alternating optimization framework and constrained concave-convex procedure technique is proposed to optimize subchannel assignment and power allocation alternately. Simulation results show that the proposed scheme outperforms the orthogonal multiple access solution and three NOMA based benchmark schemes in terms of QoE performance.