A LEO satellite routing method based on incremental evolutionary graph reinforcement learning
IEEE International Conference on Communications (ICC 2025), Montreal, Canada, Jun 8, 2025
DOI: 10.1109/ICC52391.2025.11161351
Summary in three lines
- DGA-IES addresses routing in low-Earth-orbit satellite networks with changing topology and load.
- Graph attention captures connectivity and resource state, and an incremental evolutionary strategy updates the routing policy.
- The reported comparisons show improvements in end-to-end latency, packet loss and throughput.
Abstract
Only the chairs can edit With the advent of sixth-generation (6G) technologies and growing communication demands, Low Earth Orbit (LEO) satellite networks have become essential in modern communications. However, due to the dynamic topology and complex network state of LEO environments, existing routing methods often fail to make effective decisions, limiting transmission performance. This paper proposes a novel routing method, DGA-IES. To address network state perception challenges, we introduce a topological learning model using deep graph attention (DGA), which captures complex inter-satellite connectivity and resource states. Additionally, by integrating incremental evolution strategies (IES) into deep reinforcement learning (DRL), we replace sequential interactive proximal policy optimization (PPO) with global parallel ES, achieving efficient routing convergence in the highly dynamic LEO environment. Experimental results demonstrate that our DGA-IES approach enhances LEO network load balancing by reducing end-to-end (E2E) network latency, decreasing packet loss and improving throughput compared with the benchmark approaches.
How to cite
Z. Rao, Z. Zhu, W. Y. B. Lim, Y. Yao, Y. Xu, M. Tsukada, and Y. Cheng, “A LEO satellite routing method based on incremental evolutionary graph reinforcement learning,” in IEEE International Conference on Communications (ICC 2025), Montreal, Canada, 2025, doi: 10.1109/ICC52391.2025.11161351.
BibTeX
@inproceedings{Rao2025,
title = {A LEO satellite routing method based on incremental evolutionary graph reinforcement learning},
author = {Zheheng Rao and Zhenyu Zhu and Wei Yang Bryan Lim and Ye Yao and Yanyan Xu and Manabu Tsukada and Yanyu Cheng},
doi = {10.1109/ICC52391.2025.11161351},
year = {2025},
date = {2025-06-08},
urldate = {2025-06-08},
booktitle = {IEEE International Conference on Communications (ICC 2025)},
address = {Montreal, Canada},
abstract = {Only the chairs can edit With the advent of sixth-generation (6G) technologies and growing communication demands, Low Earth Orbit (LEO) satellite networks have become essential in modern communications. However, due to the dynamic topology and complex network state of LEO environments, existing routing methods often fail to make effective decisions, limiting transmission performance. This paper proposes a novel routing method, DGA-IES. To address network state perception challenges, we introduce a topological learning model using deep graph attention (DGA), which captures complex inter-satellite connectivity and resource states. Additionally, by integrating incremental evolution strategies (IES) into deep reinforcement learning (DRL), we replace sequential interactive proximal policy optimization (PPO) with global parallel ES, achieving efficient routing convergence in the highly dynamic LEO environment. Experimental results demonstrate that our DGA-IES approach enhances LEO network load balancing by reducing end-to-end (E2E) network latency, decreasing packet loss and improving throughput compared with the benchmark approaches.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}