QoS Guaranteed Power Minimization and Beamforming for IRS-Assisted NOMA Systems

Abstract

In this letter, to solve the problem of large interference among users and reduce the transmit power consumption of base station while guaranteeing target quality of service (QoS) in downlink intelligent reflecting surface (IRS) assisted non-orthogonal multiple access (NOMA) systems, we present a joint optimization algorithm to design active beamforming for base station and passive beamforming (phase shift matrix) for IRS based on semidefinite relaxation (SDR). The non-convex constraint of signal-to-interference-plus-noise ratio (SINR) is transformed into affine constraint by increasing the dimension of optimization variables. The two subproblems of transmit power minimization and phase shift feasibility are solved iteratively using alternating optimization. To solve the computing performance degradation of SDR in largescale problems, we further propose a low-complexity algorithm based on successive convex approximation (SCA) and the original problem is then determined iteratively by relaxing the constant modulus constraint of IRS. Simulation results show that both algorithms have lower power consumption than the existing algorithms with different number of IRS reflecting elements or transmit antennas, and the SCA algorithm can approach the performance of SDR with lower complexity.