Resource Allocation and Beamforming in FIM-Assisted BS and STAR-BD-RIS-Aided NOMA: A Meta-Learning Approach

Abstract: This study explores a flexible intelligent metasurface (FIM)-based wireless communication system that integrates simultaneously transmitting and reflecting beyond diagonal reconfigurable intelligent surfaces (STAR-BD-RIS) with non-orthogonal multiple access (NOMA). The system features a multi-antenna FIM-assisted base station (BS) aided by dual-sector BD-RIS. The FIM consists of cost-effective radiating elements that can independently emit signals and dynamically adjust their vertical positions ("morphing"). The goal is to maximize energy efficiency by jointly optimizing BS beamforming, the STAR-BD-RIS matrix, NOMA constraints, and the FIM surface shape under power limits. Due to the problem's non-convexity, a meta-soft actor-critic (Meta-SAC) algorithm is proposed for adaptive optimization. Simulation results show that Meta-SAC outperforms the Meta-DDPG algorithm, and FIM-assisted designs yield substantial energy efficiency gains over benchmark schemes.