Towards Sustainable LEO Satellite Operations: A Degradation-Aware Scheduling Framework for DoD Control and Energy Optimization

Next-generation low earth orbit (LEO) satellite constellations demand the seamless integration of communication, sensing, and computation capabilities under stringent energy, thermal, and longevity constraints. This paper proposes an optimization framework that couples adaptive satellite selection with a sequential Lyapunov-based resource allocation to enable sustainable constellation operations. Deployed in a hybrid centralized-distributed architecture, the framework performs satellite selection at ground control stations while enabling autonomous onboard resource optimization.The adaptive selection mechanism uses diversity-weighted scoring across six real-time metrics, including interference awareness, for improved load balancing and interference-aware task assignment. A sequential optimization approach first derives a closed-form solution for CPU frequency under thermal and timing constraints, followed by a convex formulation for transmit power. This is integrated with a Lyapunov drift-plus-penalty framework that manages a sophisticated battery pack system using discrete binary cell-level assignment for both charging and discharging. This degradation-aware system autonomously maintains the depth-of-discharge (DoD) below 25% per cell, enforces thermal safety margins, and guarantees queue stability, ensuring long-term operational integrity. The overall problem is decomposed into polynomial-time stages suitable for real-time deployment. Simulation results demonstrate that the proposed framework reduces energy consumption by up to 84%, achieves an average DoD of 22.23% with all cells remaining below the 25% threshold, and extends the battery’s expected cycle life by 28.6% compared to baseline methods.