Deep Reinforcement Learning-Based Precoding for Multi-RIS-Aided Multiuser Downlink Systems With Practical Phase Shift
Po-Heng Chou, Boren Zheng, Wan-Jen Huang, Walid Saad, Yu Tsao, Ronald Y. Chang
Abstract
This letter considers multiple reconfigurable intelligent surfaces (RISs)-aided multiuser downlink systems with the goal of optimizing the transmitter precoding and RIS phase shift matrix to maximize spectrum efficiency. A practical coupling effect is considered between the reflecting amplitude and phase shift for the reflecting elements and their corresponding reflectivities. However, with coupled optimization variables, the formulated optimization problem is non-convex. Therefore, deep deterministic policy gradient (DDPG) based deep reinforcement learning (DRL) is proposed to solve the optimization problem. A practical scenario was simulated, whereby the proposed DDPG models were trained considering the random and fixed number of users. The simulation results show that the proposed DDPG, despite being comparably complex, outperforms the two optimization-based algorithms.