Multi-Tree Genetic Programming with Elite Recombination for dynamic task scheduling of satellite edge computing

Created by W.Langdon from gp-bibliography.bib Revision:1.8414

@Article{Zhang:2025:future,

• author = "Changzhen Zhang and Jun Yang",
• title = "Multi-Tree Genetic Programming with Elite Recombination for dynamic task scheduling of satellite edge computing",
• journal = "Future Generation Computer Systems",
• year = "2025",
• volume = "166",
• pages = "107700",
• keywords = "genetic algorithms, genetic programming, Satellite edge computing, Dynamic task scheduling, Task success rate, Elite recombination",
• ISSN = "0167-739X",
• URL = "https://www.sciencedirect.com/science/article/pii/S0167739X24006642",
• DOI = "doi:10.1016/j.future.2024.107700",
• abstract = "Satellite Edge Computing (SEC) can provide task computation services to terrestrial users, particularly in areas lacking terrestrial network coverage. With the increasing frequency of computational demands from Internet of Things (IoT) devices and the limited and dynamic nature of computational resources in Low Earth Orbit (LEO) satellites, making effective real-time scheduling decisions in dynamic environments to ensure high task success rate is a critical challenge. In this work, we investigate the dynamic task scheduling of SEC based on Genetic Programming Hyper-Heuristic (GPHH). Firstly, a new problem model for the dynamic task scheduling of SEC is proposed with the objective of improving the task success rate, where the real-world situations (limited and dynamic nature of satellite resources, randomness and difference of tasks) are taken into account. Secondly, to make efficient real-time routing decision and queuing decision during the dynamic scheduling process, a novel scheduling heuristic with routing rule and queuing rule is developed, considering dynamic features of the SEC system such as real-time load, energy consumption, and remaining deadlines. Thirdly, to automatically learn both routing rule and queuing rule, and improve the performance of the algorithm, a Multi-Tree Genetic Programming with Elite Recombination (MTGPER) is proposed, which exploits the recombination of the excellent rules to obtain the better scheduling heuristics. The experimental results show that the proposed MTGPER significantly outperforms existing state-of-the-art methods. The scheduling heuristic evolved by MTGPER has quite good interpretability, which facilitates scheduling management in engineering practice",

}

Genetic Programming entries for Changzhen Zhang Jun Yang

Citations