《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (6): 1900-1909.DOI: 10.11772/j.issn.1001-9081.2022050721

• 网络与通信 • 上一篇    下一篇

基于联邦学习的无线通信流量预测

林尚静1(), 马冀2, 庄琲1, 李月颖1, 李子怡1, 李铁1, 田锦2   

  1. 1.安全生产智能监控北京市重点实验室(北京邮电大学),北京 100876
    2.金陵科技学院 网络与通信工程学院,南京 211169
  • 收稿日期:2022-05-20 修回日期:2022-07-29 接受日期:2022-08-04 发布日期:2023-06-08 出版日期:2023-06-10
  • 通讯作者: 林尚静
  • 作者简介:林尚静(1986—),女,湖北武汉人,讲师,博士,主要研究方向:深度学习、分布式算法、边缘计算Email:linshangjing@bupt.edu.cn
    马冀(1982—),男,河北石家庄人,讲师,博士,主要研究方向:深度学习、分布式算法、边缘计算
    庄琲(2000—),女,山东菏泽人,硕士研究生,主要研究方向:大数据、边缘计算
    李月颖(1998—),女,江苏南通人,硕士研究生,主要研究方向:大数据、边缘计算
    李子怡(2000—),女,河北保定人,硕士研究生,主要研究方向:无线通信网络
    李铁(1999—),男,河北邯郸人,硕士研究生,主要研究方向:无线通信网络
    田锦(1963—),男,四川宜宾人,教授,博士,主要研究方向:人工智能算法、马尔可夫理论应用、无线网络媒体接入控制层协议、智能交通系统。
  • 基金资助:
    国家重点研发计划项目(2019YFC1511400);北京邮电大学中央高校基本科研业务费新进教师人才项目(2021RC07);泛网无线通信教育部重点实验室开放基金资助项目(KFKT-2020102);北京邮电大学研究生创新创业资助项目(2022-YC-A086)

Wireless traffic prediction based on federated learning

Shangjing LIN1(), Ji MA2, Bei ZHUANG1, Yueying LI1, Ziyi LI1, Tie LI1, Jin TIAN2   

  1. 1.Beijing Key Laboratory of Work Safety Intelligent Monitoring (Beijing University of Posts and Telecommunications),Beijing 100876,China
    2.School of Networking and Communication Engineering,Jinling Institute of Technology,Nanjing Jiangsu 211169,China
  • Received:2022-05-20 Revised:2022-07-29 Accepted:2022-08-04 Online:2023-06-08 Published:2023-06-10
  • Contact: Shangjing LIN
  • About author:MA Ji, born in 1982, Ph. D., lecturer. His research interests include deep learning, distributed algorithm, edge computing.
    ZHUANG Bei, born in 2000, M. S. candidate. Her research interests include big data, edge computing.
    LI Yueying, born in 1998, M. S. candidate. Her research interests include big data, edge computing.
    LI Ziyi, born in 2000, M. S. candidate. Her research interests include wireless communication network.
    LI Tie, born in 1999, M. S. candidate. His research interests include wireless communication network.
    TIAN Jin, born in 1963, Ph. D., professor. His research interests include artificial intelligence algorithm, application of Markov theory, wireless network media access control layer protocol, intelligent traffic system.
  • Supported by:
    National Key Research and Development Program of China(2019YFC1511400);Fundamental Research Funds for the Central Universities (BUPT)(2021RC07);Open Fund for Key Laboratory of Universal Wireless Communication, Ministry of Education(KFKT-2020102);BUPT Postgraduate Innovation and Entrepreneurship Program(2022-YC-A086)

摘要:

无线通信网络流量预测对运营商建设网络、管理基站无线资源和提升用户体验具有重要意义。然而,现有的集中式算法模型面临着复杂性和时效性问题,难以满足城市全域尺度的通信流量预测需求。因此,提出一个分布式的云边协同下的无线通信流量预测框架,以较低的复杂度和通信开销实现基于单栅格基站的流量预测。在分布式架构的基础上,提出了基于联邦学习的无线通信流量预测模型。各个栅格流量预测模型同步训练,通过中心云服务器利用JS(Jensen-Shannon)散度挑选出流量分布相似的栅格流量模型,并采用联邦平均(FedAvg)算法融合具有相似流量分布的栅格流量模型的参数,从而在提高模型泛化性的同时达到保持对本地流量精准刻画的目的。此外,由于城市范围内不同地区流量具有高度差异化的特征,在FedAvg的基础上,提出了基于合作博弈的联邦训练方法,将栅格作为合作博弈的参与者,结合超可加性准则筛选栅格,并引入合作博弈的核和夏普利值分配收益以确保联盟的稳定性,提高模型预测的准确性。实验结果表明,以短消息业务(SMS)流量为例,与栅格独立式训练相比,所提模型的预测误差下降在郊区最为明显,下降范围在26.1%~28.7%,在市区下降范围在0.7%~3.4%,在市中心下降范围在0.8%~4.7%;与栅格集中式训练相比,3个区域的模型预测误差下降范围在49.8%~79.1%。

关键词: 联邦学习, 云边协同, 无线流量预测, JS散度, 合作博弈

Abstract:

Wireless communication network traffic prediction is of great significance to operators in network construction, base station wireless resource management and user experience improvement. However, the existing centralized algorithm models face the problems of complexity and timeliness, so that it is difficult to meet the traffic prediction requirements of the whole city scale. Therefore, a distributed wireless traffic prediction framework under cloud-edge collaboration was proposed to realize traffic prediction based on single grid base station with low complexity and communication overhead. Based on the distributed architecture, a wireless traffic prediction model based on federated learning was proposed. Each grid traffic prediction model was trained synchronously, JS (Jensen-Shannon) divergence was used to select grid traffic models with similar traffic distributions through the center cloud server, and Federated Averaging (FedAvg) algorithm was used to fuse the parameters of the grid traffic models with similar traffic distributions, so as to improve the model generalization and describe the regional traffic accurately at the same time. In addition, as the traffic in different areas within the city was highly differentiated in features, on the basis of the algorithm, a federated training method based on coalitional game was proposed. Combined with super-additivity criteria, the grids were taken as participants in the coalitional game, and screened. And the core of the coalitional game and the Shapley value were introduced for profit distribution to ensure the stability of the alliance, thereby improving the accuracy of model prediction. Experimental results show that taking Short Message Service (SMS) traffic as an example, compared with grid-independent training, the proposed model has the prediction error decreased most significantly in the suburb, with a decline range of 26.1% to 28.7%, the decline range is 0.7% to 3.4% in the urban area, and 0.8% to 4.7% in the downtown area. Compared with the grid-centralized training, the proposed model has the prediction error in the three regions decreased by 49.8% to 79.1%.

Key words: federated learning, cloud-edge collaboration, wireless traffic prediction, JS (Jensen-Shannon) divergence, coalitional game

中图分类号: