Optimizing Age of Information in RIS-Assisted NOMA Networks: A Deep Reinforcement Learning Approach

Abstract

Due to the rapid development of the Internet of Things (IoT), data freshness has become particularly important. In this letter, we study a reconfigurable intelligent surface (RIS) assisted non-orthogonal multiple access (NOMA) network for collecting packets of IoT devices. Specifically, we establish a novel age of information (AoI) model to evaluate the freshness of packets. To minimize the average peak AoI, we formulate an optimization problem of jointly optimizing the phase-shift matrix of RIS and service time of packets. Then, we adopt deep deterministic policy gradient (DDPG) to solve the non-convex problem, which can handle a mass of continuous high-dimensional variables. Extensive simulation results demonstrate the superiority of the proposed scheme compared to the conventional schemes.