Wireless Power Control Based on Large Language Models

TL;DR

This paper introduces PC-LLM, a physics-informed large language model framework for wireless power control that outperforms traditional methods and neural networks, with strong zero-shot generalization and reduced inference costs.

Contribution

The paper proposes PC-LLM, integrating physical channel information into pre-trained LLMs for efficient, accurate wireless power control without retraining from scratch.

Findings

PC-LLM outperforms traditional optimization and GNN baselines.

It exhibits exceptional zero-shot generalization to unseen environments.

A lightweight adaptation reduces model depth by 50%, lowering inference costs.

Abstract

This paper investigates the power control problem in wireless networks by repurposing pre-trained large language models (LLMs) as relational reasoning backbones. In hyper-connected interference environments, traditional optimization methods face high computational cost, while standard message passing neural networks suffer from aggregation bottlenecks that can obscure critical high-interference structures. In response, we propose PC-LLM, a physics-informed framework that augments a pre-trained LLM with an interference-aware attention bias. The proposed bias tuning mechanism injects the physical channel gain matrix directly into the self-attention scores, enabling explicit fusion of wireless topology with pre-trained relational priors without retraining the backbone from scratch. Extensive experiments demonstrate that PC-LLM consistently outperforms both traditional optimization methods and state-of-the-art graph neural network baselines, while exhibiting exceptional zero-shot generalization to unseen environments. We further observe that topology-relevant relational reasoning is concentrated in shallow layers, whereas deeper layers encode task-irrelevant semantic noise. Motivated by this finding, we develop a lightweight adaptation strategy that reduces model depth by 50%, significantly lowering inference cost while preserving state-of-the-art spectral efficiency.