An Evolvable Swarm-Parallel Algorithm Framework for Satellite-Ground Networking Problem

Cheng Chen, Yonghao Du, Feng Yao2024 · DOI

Abstract

Satellite Internet has been deployed rapidly, leading to an explosive growth in satellite scale, further strengthening the contradiction between the growing demand for satellite-ground communication and the limited ground station resources. To address this challenge, this paper introduces an evolvable swarm-parallel algorithm framework (ESPAF), which comprises a swarm-parallel solving and deconflicting framework, a learning-assisted metaheuristic (LM) fast-solving algorithm that combines reinforcement learning (RL) and Tabu simulated annealing (TSA), and a linear programming (LP) exact-solving algorithm. The ESPAF breaks through the traditional task-driven satellite scheduling approach found in existing research and presents a novel and appropriate method for optimizing satellite-ground networking under a new macro paradigm. A series of experimental instances involving up to 1,000 satellites and 100 ground stations demonstrate the efficient performance of ESPAF. Comparative experiments with alternative metaheuristic algorithms and the CPLEX solver further emphasized ESP AF's ability to generate high-quality solutions more rapidly.
BibTeX
@misc{chen2024evolvable,
  author = {Cheng Chen and Yonghao Du and Feng Yao},
  title = {{An Evolvable Swarm-Parallel Algorithm Framework for Satellite-Ground Networking Problem}},
  year = {2024},
  doi = {10.1109/isaes61964.2024.10751641},
}

Loading related content...

Figure