融合可重构智能表面和深度强化学习的波束成形算法研究  

Research on Beamforming Algorithms Incorporating Reconfigurable Intelligent Surface and Deep Reinforcement Learning

作　　者：牙韩耀 万海斌[1,2] 覃团发 YA Hanyao;WAN Haibin;QIN Tuanfa(School of Computer and Electronic Information,Guangxi University,Nanning 530004,China;Guangxi Key Laboratory of Multimedia Communications and Network Technology,Guangxi University,Nanning 530004,China)

机构地区：[1]广西大学计算机与电子信息学院,南宁530004 [2]广西多媒体通信与网络技术重点实验室,南宁530004

出　　处：《小型微型计算机系统》2024年第6期1311-1317,共7页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61961004)资助。

摘　　要：可重构智能表面(Reconfigurable Intelligent Surface,RIS)技术是一种新兴的、绿色的技术,可以有效地实现频谱和能量的高效无线通信.本文研究基站(Base Stations,BSs)与RIS的联合波束形成中总发射功率最小化问题.由于人工智能(Artificial Intelligence,AI)在处理高维数据问题和非凸模型方面的优势,本文基于深度强化学习(Deep Reinforcement Learning,DRL)中的双延迟深度确定性策略梯度算法(Twin Delayed Deep Deterministic policy gradient,TD3),设计一种新颖的联合波束成形算法来处理RIS辅助无线通信系统的基站总发射功率最小化问题.仿真结果表明,本文所提算法的性能优于经典的交替优化算法,并且通过本文所提算法训练得到的模型可以直接部署和调用,不需要再次重复计算.The emerging reconfigurable intelligent surface(RIS)technology is a green technology that enables efficient wireless communication in terms of spectrum and energy.In this paper,the problem of minimizing the total transmit power in the joint beamforming of RIS and base station is studied.Due to the excellent performance of artificial intelligence(AI)techniques in addressing high-dimensional data issues and non-convex models,in this paper,based on the twin delayed deep definite policy gradient(TD3)algorithm in deep reinforcement learning(DRL),a novel joint beamforming algorithm is designed to minimize the total transmit power of base stations for RIS-assisted wireless communication systems.Simulation results reveal that the performance of the proposed algorithm is better than the classical alternating optimization algorithm,and the proposed algorithm can be directly deployed and invoked without the need to repeat the calculation.

关 键 词：可重构智能表面 深度强化学习 联合波束成形 传输功率最小化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]