Greedy and Transformer-Based Multi-Port Selection for Slow Fluid Antenna Multiple Access

TL;DR

This paper introduces two novel port-selection strategies for fluid antenna multiple access systems, combining a greedy algorithm and a Transformer-based neural network to improve spectral efficiency efficiently.

Contribution

It presents a greedy forward-selection method with swap refinement and a Transformer-based neural network trained via imitation learning and policy gradient, advancing port-selection performance and efficiency.

Findings

GFwd+S outperforms existing schemes in spectral efficiency.

Transformer-based approach approaches GFwd+S performance with lower complexity.

Proposed methods balance performance and computational cost effectively.

Abstract

We address the port-selection problem in fluid antenna multiple access (FAMA) systems with multi-port fluid antenna (FA) receivers. Existing methods either achieve near-optimal spectral efficiency (SE) at prohibitive computational cost or sacrifice significant performance for lower complexity. We propose two complementary strategies: (i) GFwd+S, a greedy forward-selection method with swap refinement that consistently outperforms state-of-the-art reference schemes in terms of SE, and (ii) a Transformer-based neural network trained via imitation learning followed by a Reinforce policy-gradient stage, which approaches GFwd+S performance at lower computational cost.