《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (11): 3616-3624.DOI: 10.11772/j.issn.1001-9081.2022111749

• 前沿与综合应用 • 上一篇    

基于图动态注意力网络的多站点风速预测

李博录, 吴利, 王晓英(), 黄建强, 曹腾飞   

  1. 青海大学 计算机技术与应用系,西宁 810016
  • 收稿日期:2022-11-24 修回日期:2023-03-10 接受日期:2023-03-17 发布日期:2023-04-03 出版日期:2023-11-10
  • 通讯作者: 王晓英
  • 作者简介:李博录(1997—),男,甘肃天水人,硕士研究生,主要研究方向:人工智能、时空气象预测
    吴利(1992—),女,安徽铜陵人,助教,硕士,主要研究方向:人工智能、高性能计算
    王晓英(1982—),女,吉林大安人,教授,博士,主要研究方向:智能电网、高性能计算、计算机体系结构 wxy_cta@qhu.edu.cn
    黄建强(1985—),男,陕西西安人,教授,博士,主要研究方向:高性能计算、大数据处理
    曹腾飞(1987—),男,湖北钟祥人,副教授,博士,主要研究方向:智能网络优化、网络攻防。
  • 基金资助:
    国家自然科学基金资助项目(62162053);清华大学—宁夏银川水联网数字治水联合研究院横向课题(SKL?IOW?2020TC2004?01);青海省科技厅应用基础研究项目(2022?ZJ?701)

Multi-site wind speed prediction based on graph dynamic attention network

Bolu LI, Li WU, Xiaoying WANG(), Jianqiang HUANG, Tengfei CAO   

  1. Department of Computer Technology and Applications,Qinghai University,Xining Qinghai 810016,China
  • Received:2022-11-24 Revised:2023-03-10 Accepted:2023-03-17 Online:2023-04-03 Published:2023-11-10
  • Contact: Xiaoying WANG
  • About author:LI Bolu, born in 1997, M. S. candidate. His research interests include artificial intelligence, spatio-temporal meteorological prediction.
    WU Li, born in 1992, M. S., teaching assistant. Her research interests include artificial intelligence, high performance computing.
    WANG Xiaoying, born in 1982, Ph. D., professor. Her research interests include smart grid, high performance computing, computer architecture.
    HUANG Jianqiang, born in 1985, Ph. D., professor. His research interests include high performance computing, big data processing.
    CAO Tengfei, born in 1987, Ph. D., associate professor. His research interests include intelligent network optimization, network attack and defense.
  • Supported by:
    National Natural Science Foundation of China(62162053);Project of Tsinghua-Ningxia Yinchuan Joint Institute of Internet of Waters on Digital Water Governance(SKL-IOW-2020TC2004-01);Application Basic Research Project of Science and Technology Department of Qinghai Province(2022-ZJ-701)

摘要:

时空序列预测任务在交通、气象、智慧城市等领域有着广泛应用。站点风速预测作为气象预测中的主要任务之一,需要结合降水、气温等外部因素,学习不同数据的时空特征。气象站点的不规则分布和风本身的固有间歇性成为实现高精度风速预测的挑战。为考虑多站点空间分布对风速的影响以获得准确可靠的预测结果,提出一种基于图的动态转换注意力网络(Graph-DSAN)风速预测模型。首先,利用不同站点之间的距离重新构建它们的连接;其次,使用局部采样的过程建模不同采样大小的邻接矩阵,实现图卷积过程中邻居节点信息的聚合与传递;接着,将时空位置编码(STPE)处理后的图卷积结果加入动态注意力编码器(DAE)和转换注意力解码器(SAD)以实现动态注意力计算,从而提取时空相关性;最后,利用自回归的方式形成多步预测。在纽约州15个站点的风速预测实验中,将所设计模型与ConvLSTM、图多注意力网络(GMAN)、时空图卷积网络(STGCN)、动态转换注意力网络(DSAN)和时空动态网络(STDN)进行比较,Graph-DSAN的12 h预测均方根误差(RMSE)分别降低了28.2%、6.9%、27.7%、14.4%和8.9%,验证了Graph-DSAN风速预测的准确性。

关键词: 风速预测, 动态注意力网络, 图卷积, 注意力机制

Abstract:

The task of spatio-temporal sequence prediction has a wide range of applications in the fields such as transportation, meteorology and smart city. It is necessary to learn the spatio-temporal characteristics of different data with the combination of external factors such as precipitation and temperature when making station wind speed predictions, which is one of the main tasks in meteorological forecasting. The irregular distribution of meteorological stations and the inherent intermittency of the wind itself bring the challenge of achieving wind speed prediction with high accuracy. In order to consider the influence of multi-site spatial distribution on wind speed to obtain accurate and reliable prediction results, a Graph-based Dynamic Switch-Attention Network (Graph-DSAN) wind speed prediction model was proposed. Firstly, the distances between different sites were used to reconstruct the connection of them. Secondly, the process of local sampling was used to model adjacency matrices of different sampling sizes to achieve the aggregation and transmission of the information between neighbor nodes during the graph convolution process. Thirdly, the results of the graph convolution processed by Spatio-Temporal Position Encoding (STPE) were fed into the Dynamic Attention Encoder (DAE) and Switch-Attention Decoder (SAD) for dynamic attention computation to extract the spatio-temporal correlations. Finally, a multi-step prediction was formed by using autoregression. In experiments on wind speed prediction on 15 sites data in New York State, the designed model was compared with ConvLSTM, Graph Multi-Attention Network (GMAN), Spatio-Temporal Graph Convolutional Network (STGCN), Dynamic Switch-Attention Network (DSAN) and Spatial-Temporal Dynamic Network (STDN). The results show that the Root Mean Square Error (RMSE) of 12 h prediction of Graph-DSAN model is reduced by 28.2%, 6.9%, 27.7%, 14.4% and 8.9% respectively, verifying the accuracy of Graph-DSAN in wind speed prediction.

Key words: wind speed prediction, dynamic attention network, graph convolution, attention mechanism

中图分类号: