Knowledge Transfer-based Large-scale Routing Scheduling Framework for Satellite Networks

The integration of space-air-ground networks represents a cutting-edge approach to delivering long-range data transmission services. As the volume of communication demands continues to surge, the allocation of resources and the design of routing strategies in satellite networks face escalating complexity. This is due to the need to manage a vast array of communication tasks. Traditional routing methods struggle to identify optimal solutions due to the dimensionality curse. To enhance service quality, this study introduces a scalable routing framework leveraging swarm intelligence. Initially, a variable grouping technique is employed to partition the initial population into smaller sub-populations, each handling lower-dimensional variables. The framework incorporates random exchange and hybrid knowledge transfer mechanisms to facilitate the generation of offspring solutions. By extracting and leveraging the rich knowledge embedded across multiple routing processes, the proposed framework significantly boosts solution efficiency. Empirical results demonstrate that the framework adeptly manages large-scale satellite routing challenges and optimally allocates network resources, outperforming existing methodologies.