Abstract
Space-air-ground integrated edge computing (SAGI-EC) networks have emerged as a novel architecture due to seamless connectivity, flexible deployment, and scalable resources. However, existing studies mainly focus on computing strategies, while challenges such as low node cooperation willingness and poor device reliability hinder the adoption of SAGI-EC. To address these challenges, we first propose a heterogeneous SAGI-EC framework that incorporates a reputation blockchain. Then, a multi-weight subjective logic model and an adaptive interaction weight updating method are employed to achieve accurate reputation evaluation and dynamic management. In addition, a problem is formulated to minimize the weighted sum of total time delay and energy consumption. By modeling the computing process as a Markov decision process and solving it through a federated reinforcement learning (FRL) approach, we propose a space-air-ground integrated edge computing algorithm based on a reputation incentive mechanism and FRL (RFSIEC). Simulation results indicate that our proposed scheme can effectively suppress and filter unreliable devices, significantly reducing the average total time delay and energy consumption while improving the task success rate.
Funding Information
  • National Natural Science Foundation of China (62171198)
  • Major Science and Technology Project of Changchun (2024WX04)
  • Science and Technology Development Project of Jilin Province (20250201082GX)

This publication has 45 references indexed in Scilit: