Energy-Efficient Resource Allocation and Antenna Selection in STAR-RIS Enabled Wireless Networks Using Meta-Learning Techniques

This paper focuses on a wireless network that utilizes simultaneously transmitting and reflecting intelligent reflecting surfaces (STAR-IRS). In this network, multiple STAR-IRSs assist a multi-antenna base station (BS) with simultaneously transmitting and reflecting signals to some single-antenna users. The goal is to maximize energy efficiency by jointly optimizing beamforming at the BS, RISs' phase shifts, and antenna selection under a maximum power budget constraint at the BS. The formulated problem is non-convex and challenging to be solved optimally. To address this difficulty, we propose a meta-deep deterministic policy gradient (Meta-DDPG) algorithm, which enables the BS to adjust its beamforming capabilities and STAR-IRSs' phase shifts and assign antennas to users. Simulation results demonstrate the superiority of Meta-DDPG in comparison with conventional DDPG. In addition, simulation results show that the multi-STAR-IRS system reaches a higher energy efficiency and data rate compared to a conventional multi- and single-IRS system.

کلیدواژه ها:

Simultaneously transmitting and reflecting intelligent reflecting surface (STAR-IRS) ، meta-learning ، deep deterministic policy gradient (DDPG) ، resource management ، antenna selection