UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

LLM-Based Port Selection and Beamforming for Multiuser MISO With Fluid Antenna Systems

Guo, Wei; Liang, Kai; Zheng, Gan; Chu, Xiaoli; Zhou, Guorong; Wong, Kai-Kit; Chae, Chan-Byoung; (2025) LLM-Based Port Selection and Beamforming for Multiuser MISO With Fluid Antenna Systems. IEEE Transactions on Network Science and Engineering pp. 1-17. 10.1109/TNSE.2025.3626002. (In press). Green open access

[thumbnail of V4_TNSE-2025_LLMFAS.pdf]
Preview
Text
V4_TNSE-2025_LLMFAS.pdf - Accepted Version

Download (774kB) | Preview

Abstract

Fluid antenna systems (FAS) introduce additional spatial degrees of freedom (DoF) by dynamically adjusting antenna positions. When integrated with multiple-input multiple-output (MIMO) technologies (forming MIMO-FAS), this capability significantly enhances wireless communication performance. However, the joint optimization of high-dimensional port selection and beamforming in MIMO-FAS presents a challenging nonconvex combinatorial problem. In this paper, we propose a novel large language model (LLM)-based intelligent framework for jointly optimizing port selection and beamforming in multiuser MIMO-FAS systems. The objective is to maximize the sum rate under base station (BS) power and port activation constraints. Departing from the conventional two-stage approaches, where port selection and beamforming are handled sequentially, we adopt a parallel output strategy that simultaneously determines port indices and beamforming coefficients, leveraging the multi-task learning capabilities of LLMs. To enhance efficiency, we incorporate low-rank adaptation (LoRA) for fine-tuning pre-trained LLMs, significantly reducing training cost while maintaining generalization performance. Also, we employ Gumbel-Sinkhorn stochastic relaxation to make discrete port selection differentiable, enabling end-to-end optimization. Numerical results demonstrate that the proposed method outperforms state-of-the-art techniques in terms of sum rate, validating the effectiveness of the LLM-driven joint optimization approach.

Type: Article
Title: LLM-Based Port Selection and Beamforming for Multiuser MISO With Fluid Antenna Systems
Open access status: An open access version is available from UCL Discovery
DOI: 10.1109/TNSE.2025.3626002
Publisher version: https://doi.org/10.1109/tnse.2025.3626002
Language: English
Additional information: This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions.
Keywords: Fluid antenna system, large language model (LLM), multiuser MISO, port selection, LoRA fine-tuning
UCL classification: UCL
UCL > Provost and Vice Provost Offices > UCL BEAMS
UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Electronic and Electrical Eng
URI: https://discovery.ucl.ac.uk/id/eprint/10217540
Downloads since deposit
0Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item