Routing of LEO Satellite Networks Based on Characteristic Graph Embedding Heuristic Approach

Low Earth Orbit (LEO) satellite networks have emerged as a pivotal component of sixth-generation (6 G) communication systems, offering substantial advantages in latency reduction, extensive coverage, and enhanced reliability. Nevertheless, their inherent heterogeneity, dynamic topologies, and frequent inter-satellite link handovers present significant challenges to conventional routing approaches. To overcome these challenges, this paper proposes an intelligent routing optimization framework, the Characteristic Graph Embedding Heuristic Approach (CGEHA). This method utilizes the Node2vec model with graph feature information to learn the topology of LEO satellite networks, generating low-dimensional embedding vectors that capture the node-edge structure and the relationship of link information, thereby effectively simplifying the complexity of satellite networks topologies. A novel node-edge feature information walk strategy enhances structural representation capability, enabling precise capture of node-edge relationships. Furthermore, the graph embedding Node2vec model with graph feature information is co-optimized with the heuristic algorithm, which significantly improves the path search efficiency and optimizes the end-to-end data transmission performance. Simulation results demonstrate that, compared to baseline algorithms, the proposed CGEHA routing method effectively reduces end-to-end delay and packet loss rate while increasing throughput, which verifies the effectiveness and availability of the method.