PPO-Based HAP Deployment for Throughput Maximization in SAGIN

Space-Air-Ground Integrated Networks (SAGINs) have attracted extensive attention as an effective solution for providing worldwide communication services and supporting diverse transmission requirements. Nevertheless, the highly dynamic wireless environment and fluctuating user distribution introduce considerable difficulties for efficient deployment of High-Altitude Platforms (HAPs). This work studies a throughput enhancement problem in a SAGIN architecture composed of a Low Earth Orbit (LEO) satellite, multiple HAPs, and terrestrial users. Considering practical communication constraints such as beam scheduling and subchannel assignment, the objective is to improve the overall data transmission capability through optimized HAP positioning. By interacting with varying network states, the proposed learning framework agent continuously learns suitable deployment policies for HAPs, thereby improving transmission performance and spectrum utilization efficiency. Numerical evaluations indicate that the policy optimization strategy achieves superior throughput performance and communication efficiency compared with traditional fixed deployment approaches.