Deep neural network based channel allocation for interference-limited wireless networks

Zhenfeng Zhang, Daosen Zhai, Ruonan Zhang, Yutong Wang

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

4 Scopus citations

Abstract

Cooperative communication in wireless networks has received much attention in both academia and industry. How to effectively allocate and schedule radio resources to improve system performance becomes an important issue of cooperative communication. This paper mainly studies the ultra-low complexity wireless channel allocation algorithm for interference-limited networks. Firstly, we use the traditional sequential convex approximation (SCA) technique to design the channel allocation algorithm. Then, we utilize the characteristics of deep neural network (DNN) that can approximate a complex function with multiple layers of mapping to approximate the SCA-based algorithm. Based on DNN, we design an ultra-low complexity algorithm. Simulation results indicate that the DNN-based algorithm can achieve good performance with ultra-low computation time, which is a feature for practical application.

Original languageEnglish
Title of host publication2019 IEEE 20th International Conference on High Performance Switching and Routing, HPSR 2019
PublisherIEEE Computer Society
ISBN (Electronic)9781728116860
DOIs
StatePublished - May 2019
Event20th IEEE International Conference on High Performance Switching and Routing, HPSR 2019 - Xi'An, China
Duration: 26 May 201929 May 2019

Publication series

NameIEEE International Conference on High Performance Switching and Routing, HPSR
Volume2019-May
ISSN (Print)2325-5595
ISSN (Electronic)2325-5609

Conference

Conference20th IEEE International Conference on High Performance Switching and Routing, HPSR 2019
Country/TerritoryChina
CityXi'An
Period26/05/1929/05/19

Keywords

  • cooperative communication
  • deep neural network
  • sequential convex approximation
  • wireless channel allocation

Fingerprint

Dive into the research topics of 'Deep neural network based channel allocation for interference-limited wireless networks'. Together they form a unique fingerprint.

Cite this