Sum-Rate Maximization for RIS-Aided Full-Duplex Non-Orthogonal Multiple Access Networks: Joint Passive Beamforming and Resource Allocation

Abstract

This study investigates the resource allocation problem of a reconfigurable intelligent surface (RIS)-aided full-duplex (FD) multi-channel (MC) non-orthogonal multiple access (NOMA) networks. The weighted sum-rate maximization problem is formulated to jointly optimization of sub-channel assignment, decoding order of downlink (DL) and uplink (UL) users, UL transmit power, transmit and receive beamforming vectors at the BS, and reflection coefficients of the RIS. To address the highly non-convex problem, an alternating optimization (AO) method is adopted to decompose the original problem into three subproblems. First, a novel approach is proposed for jointly optimizing the decoding order, sub-channel assignment, UL transmit power, and transmit beamforming vectors by adopting a penalty-based method, majorization minimization (MM) and semi-definite relaxation approaches. Then, by employing the generalized eigenvalue problem, an optimal closed-form solution for the receive beamforming vectors is devised. Finally, an iterative algorithm based on a sequential rank-one constraint relaxation approach (SROCR) together with the MM technique is adopted to design the reflection coefficients of the RIS. Simulation results confirm that the proposed method outperforms the conventional RIS-aided MC half-duplex NOMA (MC-HD-NOMA) and MC FD orthogonal multiple access (MC-FD-OMA) networks while having comparable performance to the exhaustive search method in terms of spectral efficiency.