Efficient information support in LEO SCNs: A multi-layer spatiotemporal cost model and dynamic reallocation algorithm

To enable global coverage and massive connectivity in the 6G era, space-air-ground integrated networks (SAGINs) rely heavily on low Earth orbit (LEO) satellite constellations, which offer wide-area coverage and autonomous inter-satellite links (ISLs). However, the high mobility, dynamic topology, and uneven resource distribution in LEO networks pose significant challenges to service continuity, efficient scheduling, and resource utilization. This paper proposes an innovative multi-layer spatiotemporal cost graph model that captures the time-varying network dynamics through discretized time windows and layered node representations, transforming complex ISL scheduling into a constrained flow optimization problem. Based on this model, we design a priority-based routing (PBR) algorithm, where mission data volume serves as a priority metric for flow allocation. By periodically updating the spatiotemporal graph and re-executing PBR, the system achieves adaptive scheduling through a system-level feedback mechanism. Evaluations on the Globalstar-2, Iridium Next, and Telesat constellations show that the proposed approach improves throughput by 300%, reduces latency by 50%, and increases mission completion rates by 20%, demonstrating significant gains in inter-satellite resource efficiency.