Radio resource scheduling for 5G NR via deep deterministic policy gradient

Sheng Chia Tseng, Zheng Wei Liu, Yen Cheng Chou, Chih Wei Huang

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

13 Scopus citations

Abstract

The fifth generation (5G) wireless system plays a crucial role to realize future network applications with diverse services requirements. The 3rd Generation Partnership Project (3GPP) proposed 5G New Radio (NR) specifications with significantly greater flexibility on configurations and procedures to facilitate a more efficient and agile radio access network (RAN). At the same time, the complexity of resource management increases, and the advantage of machine learning techniques are worth studying. In this article, we investigate the radio resource scheduling issue in the 5G RAN. Through a modularized deep deterministic policy gradient (DDPG) architecture and specifically defined action as a combination of scheduling algorithms Through specifically defined action as a combination of scheduling algorithms, the proposed method is efficient to train and performing well. Favorable results are observed compared with conventional scheduling algorithms. The proposed architecture applies to other radio resource management problems with similar characteristic.

Original languageEnglish
Title of host publication2019 IEEE International Conference on Communications Workshops, ICC Workshops 2019 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728123738
DOIs
StatePublished - May 2019
Event2019 IEEE International Conference on Communications Workshops, ICC Workshops 2019 - Shanghai, China
Duration: 20 May 201924 May 2019

Publication series

Name2019 IEEE International Conference on Communications Workshops, ICC Workshops 2019 - Proceedings

Conference

Conference2019 IEEE International Conference on Communications Workshops, ICC Workshops 2019
Country/TerritoryChina
CityShanghai
Period20/05/1924/05/19

Keywords

  • 5G NR
  • Deep learning
  • Radio resource scheduling
  • Reinforcement learning

Fingerprint

Dive into the research topics of 'Radio resource scheduling for 5G NR via deep deterministic policy gradient'. Together they form a unique fingerprint.

Cite this