Hybrid Large Language Models and Reinforcement Learning for Energy-Efficient Multisatellite Scheduling: Boosting the Performance From Scratch
- 16 December 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (3) , 5379-5392
Abstract
Low Earth orbit (LEO) satellite constellations are crucial for global connectivity by providing extensive coverage and reduced delays. However, scheduling data transmission in these dynamic networks is challenging due to rapidly changing satellite positions. This research introduces boosted reinforcement learning (BoostRL), a novel framework integrating large language models (LLMs) with reinforcement learning (RL) for efficient scheduling in LEO satellite constellations. BoostRL leverages LLM-generated initial policies to accelerate convergence, thereby guiding the early stages of policy learning while adapting swiftly to dynamic network conditions. It uses a hybrid policy approach which transitions smoothly from LLM recommendations to autonomous RL policies. A tailored initialization of $Q$ -value parameters and an enhanced loss function further optimize learning efficiency, by aligning the initial learning phase with LLM-generated insights. Simulations using two-line element (TLE) orbital data demonstrate that BoostRL achieves rapid convergence and improved efficiency, thereby validating its potential as a scalable, adaptive solution for managing satellite communication networks.Keywords
Funding Information
- Intelligent Technology Development Program on Disaster Response and Emergency Management funded by the Ministry of Interior and Safety (2022-MOIS37-005)
- Ministry of Science and ICT (MSIT), Korea, under the Information Technology Research Center (ITRC) support program (IITP-2026-RA-2024-0043887)
- Institute for Information and Communications Technology Planning and Evaluation
This publication has 36 references indexed in Scilit:
- A Novel Satellite-Based REM Construction in Cognitive GEO-LEO Satellite IoT NetworksIEEE Internet of Things Journal, 2024
- Multipath Cooperative Routing in Ultradense LEO Satellite Networks: A Deep-Reinforcement-Learning-Based ApproachIEEE Internet of Things Journal, 2024
- Markov Decision Policies for Distributed Angular Routing in LEO Mobile Satellite Constellation NetworksIEEE Internet of Things Journal, 2024
- Enabling High-Throughput Routing for LEO Satellite Broadband Networks: A Flow-Centric Deep Reinforcement Learning ApproachIEEE Internet of Things Journal, 2024
- Reinforcement Learning Based Intelligent Routing for Software Defined LEO Satellite Networks2023
- Exploiting topology awareness for routing in LEO satellite constellations2021
- Deep Dyna-Reinforcement Learning Based on Random Access Control in LEO Satellite IoT NetworksIEEE Internet of Things Journal, 2021
- Max Completion Time Optimization for Internet of Things in LEO Satellite-Terrestrial Integrated NetworksIEEE Internet of Things Journal, 2021
- QoE-Aware Intelligent Satellite Constellation Design in Satellite Internet of ThingsIEEE Internet of Things Journal, 2020
- A multicast routing algorithm for LEO satellite IP networksIEEE/ACM Transactions on Networking, 2002