LLM guided scheduling boosts satellite contact time by nearly 15 percent

CIPHEUR: LLM-Guided Structural Search Control for Conflict-Graph Optimization

Neural and Evolutionary Computing

Summary

Scheduling satellites to communicate with ground stations involves complex conflicts that can stall progress. The authors introduce CIPHEUR, which uses large language models to better identify and fix these conflicts during the scheduling process. This approach leads to noticeably longer communication windows and outperforms previous automated methods. Their method balances trying new solutions with keeping the overall goals and rules intact, helping the scheduler make steady improvements.

What this means in practice

  • For satellite operations teams: Improve satellite ground contact scheduling to increase total communication time and efficiency by integrating LLM-guided interventions.
  • For industrial scheduling engineers: Enhance complex conflict-resolution in resource allocation problems by adding an LLM-driven decision layer for coordinated interventions during search.

Authors

Yuqing Cheng, Yan Jia, Yanjie Song, Xiaolu Liu, Lining Xing, Feng Yao

Abstract

Conflict-graph search can stagnate despite effective operators when interventions fail to address the conflicts blocking progress. A disruptive intervention may also require further search before its value emerges. We propose CIPHEUR, a large language model (LLM)-guided framework that connects conflict evidence to coordinated intervention and recovery through Structured Intervention Units (SIUs). The shared control interface supports offline synthesis of executable policies and online contextual decisions. Asynchronous online control sustains native search during inference, while all policies preserve the original objective and feasibility constraints. Across eight satellite-ground scheduling conflict views with five runs per view and a 900-second evaluation cap, CIPHEUR-online achieves the highest mean objective among 11 configurations and increases mean scheduled contact duration by 14.84% over the greedy baseline GWMIN. It exceeds the strongest external per-view mean in seven of eight views and outperforms CHILS in all 40 paired runs. Within the same runs, its normalized objective improvement from 360 to 900 seconds is 1.50 to 1.86 times that of the two programmatic controllers. A separately frozen, one-shot LLM-generated controller also improves the mean objective over CHILS without runtime model calls. Together, these results demonstrate how structural search control translates LLM design knowledge and contextual reasoning into improved solution quality and sustained search progress.