Low-Complexity Algorithm for Maximizing the Weighted Sum-Rate of Intelligent Reflecting Surface-Assisted Wireless Networks
Yajun Wang, Lili Fang, Shanjie Cai, Zhuxian Lian, Yinjie Su, Zhibin Xie
Abstract
Intelligent reflecting surface (IRS) via using massive low-cost passive elements that can reflect the signals by adjusting phase shifts provides a cost-effective and energy-efficient solution to enhance the wireless communication system’s performance. In the article, we consider an IRS-aided multiuser multi-input–single-output (MISO) downlink system. We tackle the weighted sum-rate (WSR) maximization by jointly optimizing the active beamforming at the base station (BS) and the passive beamforming at the IRS. We first decouple the nonconvex optimization problem by the Lagrangian dual transform, then resort to fractional programming to address the active and passive beamforming optimizations. We develop the mirror descent (MD) method and the accelerated projected gradient (APG) method to solve subproblems. The simulation results show that the MD and APG algorithm get the comparable WSR gain and convergence speeds as existing methods, but with a significantly lower computational complexity.