Minimizing Energy Consumption of Satellite Edge Computing-Capable LEO Satellites

Low Earth Orbit (LEO) satellite constellations have gained considerable attention from both academia and industry due to their potential to provide reliable global coverage. However, the continuous movement of satellites causes periodic transitions between the sunlit and shadowed regions of orbit, leading to fluctuations in available energy. Satellites on the sunlit side can harvest solar energy to recharge their batteries, while those in the shadow primarily rely on stored battery power. This paper presents a joint task and energy allocation framework that optimizes satellite energy consumption by considering three key factors: onboard processing of small tasks, relaying tasks to sunlit satellites for processing, and offloading larger tasks to ground stations when available. An Integer Linear Programming (ILP) approach is employed to determine the optimal energy distribution across these tasks, while a computationally efficient greedy algorithm is introduced as an alternative. Additionally, Dynamic Voltage and Frequency Scaling (DVFS) is incorporated to optimize the energy consumption during task processing by adjusting the processing frequency according to the available energy. The results indicate that ILP achieves optimal energy efficiency, while the greedy approach provides a near-optimal solution with just a 4.4% deviation.