Service Priority-Driven Joint Beam Hopping and Power Allocation in LEO Satellite Networks

In low earth orbit (LEO) satellite communication systems, spatiotemporal traffic non-uniformity and service heterogeneity pose significant challenges for efficient resource allocation. Existing beam hopping (BH) strategies often fail to accurately model quality of service (QoS) priorities, resulting in mismatched resource allocation. This paper introduces a priority-driven BH optimization approach that aims to maximize throughput and improve fairness by jointly considering service characteristics and dynamic traffic. A hybrid fuzzy analytic hierarchy process (FAHP) and the entropy weight method (EWM) are employed to compute static weight based on service categories and demand proportions, while a sliding window mechanism monitors traffic and queue states to derive dynamic weight. These weights are integrated into a cell-level priority index that guides beam scheduling. A deep reinforcement learning (DRL) algorithm based on proximal policy optimization (PPO) then optimizes the BH pattern and power allocation according to the computed priorities. Experimental results show that the proposed method outperforms benchmark schemes in throughput, fairness, and queuing delay, while maintaining a consistently low inter-cell delay gap even under heterogeneous traffic scenarios.