Traffic-Aware Domain Partitioning and Load-Balanced Inter-Domain Routing for LEO Satellite Networks

2026-04-14

AI Summary

This work addresses the challenges of inter-domain routing in low Earth orbit (LEO) satellite networks, where high node mobility, uneven traffic distribution, and frequent link failures lead to load imbalance and reduced reliability. To tackle these issues, the authors propose a joint optimization framework integrating multi-objective optimization with deep reinforcement learning. First, the NSGA-II algorithm is employed offline to generate domain partitions that exhibit high cohesion and low load imbalance. Then, an online routing decision mechanism is developed by combining a graph attention network (GAT)-based link state awareness module with an action-masked proximal policy optimization (PPO) agent. Notably, this is the first application of GAT and action-masked PPO to LEO inter-domain routing, enabling joint perception of dynamic topology, real-time traffic, and failure conditions. Simulations on a 288-satellite Walker constellation demonstrate significant reductions in link load imbalance and end-to-end delay, along with improved routing success rates and lower packet loss across normal, traffic surge, and failure scenarios.

Technology Category

• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Search and Optimization: Learning to Search
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Low Earth Orbit (LEO) satellite networks provide global coverage and low latency, yet high node mobility, uneven traffic distribution, and stochastic link failures pose severe challenges for inter-domain routing. Existing approaches either neglect graph-structured topology or lack dynamic awareness of real-time link states, struggling to balance load distribution and routing reliability. This paper proposes DTAR, a traffic-aware deep reinforcement learning approach for inter-domain routing in LEO satellite networks. A multi-objective NSGA-II algorithm first generates an offline domain partition maximizing intra-domain traffic ratio and minimizing load imbalance. A Graph Attention Network dynamically encodes inter-domain link traffic intensity, load distribution, and fault status, upon which an action-masked PPO agent learns routing decisions online. Simulations on a 288-satellite Walker constellation against multiple baselines demonstrate that DTAR significantly reduces link load imbalance and end-to-end delay, while improving routing success rate and reducing packet loss rate across normal, traffic surge, and fault scenarios.

Problem

• LEO satellite networks
• inter-domain routing
• load balancing
• traffic awareness
• link failures

Innovation

• Traffic-aware routing
• Domain partitioning
• Graph Attention Network
• Deep reinforcement learning
• LEO satellite networks