浙江电力

2021, v.40;No.301(05) 66-71

[打印本页] [关闭]
本期目录(Current Issue) | 过刊浏览(Archive) | 高级检索(Advanced Search)

基于多态分量与3D卷积相融合的带宽预测方法
A Bandwidth Prediction Method Based on Dynamic and Static Components and 3D Convolution Convergence

段浩,赵璞,郑思源,张清周,季节,吴昌设,张晋阳
DUAN Hao,ZHAO Pu,ZHENG Siyuan,ZHANG Qingzhou,JI Jie,WU Changshe,ZHANG Jinyang

摘要(Abstract):

随着"两网"融合的逐步推广,可靠的流量预测对于改善城市网络环境、优化网络效率起到至关重要的作用。针对目前带宽预测方法存在的带宽利用率不高、仅针对预测当地进行研究而忽视相邻地区的影响等问题,提出了一种基于3D卷积和最小二乘法的动态、静态分量的通信带宽估算方法。搭建了相对应的网络模型,根据城市网络一定时间段内的流量分布,得到城市不同范围内的通信带宽值和信息流量密度,并通过算例验证了所提方法的有效性。
With the promotion of "dual-network" convergence, reliable data prediction is playing a key role in urban network environment improvement and network efficiency optimization. In view of the low bandwidth utilization ratio of the existing bandwidth prediction method and the research focusing merely on the local areas while ignoring the impact of adjacent areas, this paper proposes a dynamic and static component communication bandwidth estimation method based on 3D convolution and least squares. A corresponding network model is built. According to the traffic distribution in a certain period of the city network, the communication bandwidth value and information flow density in different ranges of the city are obtained, and the effectiveness of the proposed method is verified by a calculation example.

关键词(KeyWords): 3D卷积;最小二乘法;静态分量;动态分量;带宽预测
3D convolution;least squares method;static component;dynamic component;bandwidth prediction

Abstract:

Keywords:

基金项目(Foundation): 国网浙江省电力有限公司温州供电公司科技项目(2020-297)

作者(Author): 段浩,赵璞,郑思源,张清周,季节,吴昌设,张晋阳
DUAN Hao,ZHAO Pu,ZHENG Siyuan,ZHANG Qingzhou,JI Jie,WU Changshe,ZHANG Jinyang

DOI: 10.19585/j.zjdl.202105010

参考文献(References):

扩展功能
本文信息
服务与反馈
本文关键词相关文章
本文作者相关文章
中国知网
分享