Large Language Models for Control

TL;DR

This paper explores using large language models to generate control actions directly, comparing different variants and demonstrating their viability for control tasks without traditional control-engineering.

Contribution

It introduces and evaluates multiple LLM-based control methods, highlighting their potential for adaptive and human-in-the-loop control in cyber-physical systems.

Findings

Prompt-only LLMs can produce viable control actions.

Tool-assisted LLMs adapt better to changing objectives.

Larger models are sensitive to constraints and actuation effort.

Abstract

This paper investigates using large language models (LLMs) to generate control actions directly, without requiring control-engineering expertise or hand-tuned algorithms. We implement several variants: (i) prompt-only, (ii) tool-assisted with access to historical data, and (iii) prediction-assisted using learned or simple models to score candidate actions. We compare them on tracking accuracy and actuation effort, with and without a prompt that requests lower actuator usage. Results show prompt-only LLMs already produce viable control, while tool-augmented versions adapt better to changing objectives but can be more sensitive to constraints, supporting LLM-in-the-loop control for evolving cyber-physical systems today and operator and human inputs.