Home > Research > Publications & Outputs > Entropy-based Reinforcement Learning for comput...

Electronic data

  • FGCS_Author manuscript

    Rights statement: This is the author’s version of a work that was accepted for publication in Future Generation Computer Systems. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Future Generation Computer Systems, 136, 2022 DOI: 10.1016/j.future.2022.06.002

    Accepted author manuscript, 674 KB, PDF document

    Available under license: CC BY-NC-ND: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License

Links

Text available via DOI:

View graph of relations

Entropy-based Reinforcement Learning for computation offloading service in software-defined multi-access edge computing

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published
  • K. Li
  • X. Wang
  • Q. Ni
  • M. Huang
Close
<mark>Journal publication date</mark>30/11/2022
<mark>Journal</mark>Future Generation Computer Systems
Volume136
Number of pages11
Pages (from-to)241-251
Publication StatusPublished
Early online date20/06/22
<mark>Original language</mark>English

Abstract

The rapid growth of Internet of Things (IoT) devices and the emergence of multiple edge applications have resulted in an explosive growth of data traffic at the edge of the networks. Computation offloading services in Multi-access edge computing (MEC) enabled networks to offer potentials of a better Quality of Service (QoS) than traditional networks. They are expected to reduce the propagation delay and enhance the computational capability for delay-sensitive tasks especially. Nevertheless, the distributed computing resources of edge devices urgently need reasonable resource controllers to ensure such distributed computing resources to be effectively scheduled. The benefits of Software-Defined Networking (SDN) may be explored to demonstrate their full potential through MEC services to reduce the response time of programs. In this paper, a new SDN-based MEC computation offloading service architecture is proposed to increase the coordination and offloading capabilities at the control plane. Besides, to deal with dynamic network changes and increase the exploration degree, we propose a novel Entropy-based Reinforcement Learning algorithm for delay-sensitive tasks computation offloading at the edge of the networks. Finally, the evaluation findings indicate that our proposed model has the potential to improve the network resource allocation and balanced performance significantly.

Bibliographic note

This is the author’s version of a work that was accepted for publication in Future Generation Computer Systems. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Future Generation Computer Systems, 136, 2022 DOI: 10.1016/j.future.2022.06.002