Zhao Di, Zheng Zhong, Pengfei Qin, Qin Hao, Song Bin
{"title":"Resource allocation in multi-user cellular networks: A transformer-based deep reinforcement learning approach","authors":"Zhao Di, Zheng Zhong, Pengfei Qin, Qin Hao, Song Bin","doi":"10.23919/JCC.ea.2021-0665.202401","DOIUrl":null,"url":null,"abstract":"To meet the communication services with diverse requirements, dynamic resource allocation has shown increasing importance. In this paper, we consider the multi-slot and multi-user resource allocation (MSMU-RA) in a downlink cellular scenario with the aim of maximizing system spectral efficiency while guaranteeing user fairness. We first model the MSMU-RA problem as a dual-sequence decision-making process, and then solve it by a novel Transformer-based deep reinforcement learning (TDRL) approach. Specifically, the proposed TDRL approach can be achieved based on two aspects: 1) To adapt to the dynamic wireless environment, the proximal policy optimization (PPO) algorithm is used to optimize the multi-slot RA strategy. 2) To avoid co-channel interference, the Transformer-based PPO algorithm is presented to obtain the optimal multi-user RA scheme by exploring the mapping between user sequence and resource sequence. Experimental results show that: i) the proposed approach outperforms both the traditional and DRL methods in spectral efficiency and user fairness, ii) the proposed algorithm is superior to DRL approaches in terms of convergence speed and generalization performance.","PeriodicalId":9814,"journal":{"name":"China Communications","volume":null,"pages":null},"PeriodicalIF":3.1000,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"China Communications","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.23919/JCC.ea.2021-0665.202401","RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"TELECOMMUNICATIONS","Score":null,"Total":0}
引用次数: 0
Abstract
To meet the communication services with diverse requirements, dynamic resource allocation has shown increasing importance. In this paper, we consider the multi-slot and multi-user resource allocation (MSMU-RA) in a downlink cellular scenario with the aim of maximizing system spectral efficiency while guaranteeing user fairness. We first model the MSMU-RA problem as a dual-sequence decision-making process, and then solve it by a novel Transformer-based deep reinforcement learning (TDRL) approach. Specifically, the proposed TDRL approach can be achieved based on two aspects: 1) To adapt to the dynamic wireless environment, the proximal policy optimization (PPO) algorithm is used to optimize the multi-slot RA strategy. 2) To avoid co-channel interference, the Transformer-based PPO algorithm is presented to obtain the optimal multi-user RA scheme by exploring the mapping between user sequence and resource sequence. Experimental results show that: i) the proposed approach outperforms both the traditional and DRL methods in spectral efficiency and user fairness, ii) the proposed algorithm is superior to DRL approaches in terms of convergence speed and generalization performance.
期刊介绍:
China Communications (ISSN 1673-5447) is an English-language monthly journal cosponsored by the China Institute of Communications (CIC) and IEEE Communications Society (IEEE ComSoc). It is aimed at readers in industry, universities, research and development organizations, and government agencies in the field of Information and Communications Technologies (ICTs) worldwide.
The journal's main objective is to promote academic exchange in the ICTs sector and publish high-quality papers to contribute to the global ICTs industry. It provides instant access to the latest articles and papers, presenting leading-edge research achievements, tutorial overviews, and descriptions of significant practical applications of technology.
China Communications has been indexed in SCIE (Science Citation Index-Expanded) since January 2007. Additionally, all articles have been available in the IEEE Xplore digital library since January 2013.