GUIDE: Guided Updates for In-context Decision Evolution in LLM-Driven Spacecraft Operations

TL;DR

GUIDE is a framework that improves LLM-based spacecraft control by evolving a structured decision rule playbook across episodes without retraining, leading to better performance in simulated orbital tasks.

Contribution

It introduces a novel non-parametric policy improvement method enabling cross-episode adaptation of LLM agents through structured decision rules.

Findings

GUIDE outperforms static prompting baselines in orbital interception tasks.

Context evolution acts as policy search over structured decision rules.

Real-time control is achieved with a lightweight acting model.

Abstract

Large language models (LLMs) have been proposed as supervisory agents for spacecraft operations, but existing approaches rely on static prompting and do not improve across repeated executions. We introduce \textsc{GUIDE}, a non-parametric policy improvement framework that enables cross-episode adaptation without weight updates by evolving a structured, state-conditioned playbook of natural-language decision rules. A lightweight acting model performs real-time control, while offline reflection updates the playbook from prior trajectories. Evaluated on an adversarial orbital interception task in the Kerbal Space Program Differential Games environment, GUIDE's evolution consistently outperforms static baselines. Results indicate that context evolution in LLM agents functions as policy search over structured decision rules in real-time closed-loop spacecraft interaction.