Parallel Task Offloading and Resource Optimization for LEO Satellites Edge Computing

This study examines the configuration of Low Earth Orbit (LEO) satellite networks that employ laser-based intersatellite links (LISLs) for satellite edge computing (SEC). High-resolution applications like synthetic aperture radar (SAR) imaging use LEO satellite systems, which face significant challenges in effectively regulating energy consumption and reducing work latency. To address these challenges, we proposed a novel approach that utilizes the lowest-cost method based on a genetic algorithm (LCMGA). This strategy focuses on allocating computing subtasks and optimizing energy consumption by parallel offloading them over multiple SEC nodes. The primary objective is to decrease energy costs while upholding task latency demands. The LCMGA algorithm utilizes the computing resource of edge satellites in the LISL range to determine the most efficient methods of offloading tasks. The simulation findings demonstrate that LCMGA significantly reduces the computational time for processing large volumes of data, achieving processing durations of only a few seconds. In addition, LCMGA surpasses existing algorithms in terms of energy usage and CPU energy efficiency.