基于动态时空Transformer的城市蜂窝网络流量预测方法

doi:10.3778/j.issn.1002-8331.2405-0311

摘要/Abstract

摘要： 针对现有的城市蜂窝网络流量预测方法没有考虑到小区间空间相关性的动态性以及不同时间跨度下小区间空间相关性的多样性问题，提出基于动态时空Transformer的城市蜂窝网络流量预测模型（DSTTNet）。提出多尺度时间感知空间Transformer模块MSTAST，通过分时间段建模小区间的空间相关性，实现小区间动态空间关系的捕获；通过引入多分支结构，在不同的分支上使用不同的时间段划分方式来捕获不同的空间相关性，从而提高空间相关性建模的准确性；基于MSTAST和时间Transformer模块构建时空序列建模模块来捕获城市蜂窝网络流量中的长时间依赖关系和动态空间依赖关系；还将MSTAST应用于特征融合模块，以提高模型对预测特征中远距离小区间全局空间关系的捕获能力。实验结果表明，在RMSE评价指标下，所提模型在SMS、Call、Internet三种网络流量数据集上分别提升了5.43%、4.30%、2.86%。

关键词: 蜂窝网络流量预测, 时空数据挖掘, Transformer, 注意力机制, 无线网络

Abstract: Existing citywide cellular traffic prediction methods fail to consider the dynamic nature of spatial correlations between cells, as well as the diversity of spatial correlations. To address these issues, this research introduces a citywide cellular traffic prediction model named dynamic spatial temporal transformer network (DSTTNet). Initially, a multi-scale time aware spatial Transformer module (MSTAST) is proposed to capture dynamic spatial relationships by modeling the spatial correlation in different time periods. By introducing a multi-branch structure, different time period divisions are utilized on different branches to capture diverse spatial correlations. The features from different branches complement each other, thereby enhancing the accuracy of spatial correlation modeling. Subsequently, a spatiotemporal series modeling module, based on MSTAST and temporal Transformer module, is constructed to capture long-term dependencies and dynamic spatial dependencies in citywide cellular traffic. Furthermore, MSTAST is integrated into the feature fusion module to enhance the model’s capability in capturing global spatial relationships between distant intervals in predicted features. Experimental results show that under the RMSE evaluation metric, the proposed model improves by 5.43%, 4.30% and 2.86% on the three cellular traffic datasets: SMS, Call and Internet, respectively.

Key words: citywide cellular traffic prediction, spatial temporal data mining, Transformer, attention mechanism, wireless networks

于江燕, 王倩, 孟宪静, 张瑞敏, 耿蕾蕾. 基于动态时空Transformer的城市蜂窝网络流量预测方法[J]. 计算机工程与应用, 2025, 61(18): 290-299.

YU Jiangyan, WANG Qian, MENG Xianjing, ZHANG Ruimin, GENG Leilei. Dynamic Spatial Temporal Transformer for Citywide Cellular Traffic Prediction[J]. Computer Engineering and Applications, 2025, 61(18): 290-299.

参考文献

[1] LIN J S, CHEN Y J, ZHENG H F, et al. A data-driven base station sleeping strategy based on traffic prediction[J]. IEEE Transactions on Network Science and Engineering, 2024, 11(6): 5627-5643.
[2] 康梦轩, 宋俊平, 范鹏飞, 等. 基于深度学习的网络流量预测研究综述[J]. 计算机工程与应用, 2021, 57(10): 1-9.
KANG M X, SONG J P, FAN P F, et al. Survey of network traffic forecast based on deep learning[J]. Computer Engineering and Applications, 2021, 57(10): 1-9.
[3] JIANG N, DENG Y S, NALLANATHAN A. Traffic prediction and random access control optimization: learning and non-learning-based approaches[J]. IEEE Communications Magazine, 2021, 59(3): 16-22.
[4] FERREIRA D, BRAGA REIS A, SENNA C, et al. A forecasting approach to improve control and management for 5G networks[J]. IEEE Transactions on Network and Service Management, 2021, 18(2): 1817-1831.
[5] GO?CIE? R. Efficient dynamic routing in elastic optical networks based on traffic prediction and bandwidth reservation[C]//Proceedings of the 2023 23rd International Conference on Transparent Optical Networks. Piscataway: IEEE, 2023: 1-4.
[6] ZHOU X H, BILAL M, DOU R H, et al. Edge computation offloading with content caching in 6G-enabled IoV[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(3): 2733-2747.
[7] GU B, ZHAN J H, GONG S M, et al. A spatial-temporal transformer network for city-level cellular traffic analysis and prediction[J]. IEEE Transactions on Wireless Communications, 2023, 22(12): 9412-9423.
[8] TRAN Q T, HAO L, TRINH Q K. A comprehensive research on exponential smoothing methods in modeling and forecasting cellular traffic[J]. Concurrency and Computation: Practice and Experience, 2020, 32(23): e5602.
[9] XU F L, LIN Y Y, HUANG J X, et al. Big data driven mobile traffic understanding and forecasting: a time series approach[J]. IEEE Transactions on Services Computing, 2016, 9(5): 796-805.
[10] XIA H W, WEI X, GAO Y, et al. Traffic prediction based on ensemble machine learning strategies with bagging and LightGBM[C]//Proceedings of the 2019 IEEE International Conference on Communications Workshops. Piscataway: IEEE, 2019: 1-6.
[11] 田中大, 潘信澎. 小波消噪和优化支持向量机的网络流量预测[J]. 北京邮电大学学报, 2022, 45(5): 79-84.
TIAN Z D, PAN X P. Network traffic prediction using wavelet denoising and optimized support vector machine[J]. Journal of Beijing University of Posts and Telecommunications, 2022, 45(5): 79-84.
[12] WANG J, TANG J, XU Z Y, et al. Spatiotemporal modeling and prediction in cellular networks: a big data enabled deep learning approach[C]//Proceedings of the IEEE Conference on Computer Communications. Piscataway: IEEE, 2017: 1-9.
[13] TRAN D, BOURDEV L, FERGUS R, et al. Learning spatiotemporal features with 3D convolutional networks[C]//Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2015: 4489-4497.
[14] HUANG C W, CHIANG C T, LI Q H. A study of deep learning networks on mobile traffic forecasting[C]//Proceedings of the 2017 IEEE 28th Annual International Symposium on Personal, Indoor, and Mobile Radio Communications. Piscataway: IEEE, 2017: 1-6.
[15] ZHANG C T, ZHANG H X, YUAN D F, et al. Citywide cellular traffic prediction based on densely connected convolutional neural networks[J]. IEEE Communications Letters, 2018, 22(8): 1656-1659.
[16] HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutional networks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 2261-2269.
[17] ZHANG C Y, PATRAS P. Long-term mobile traffic forecasting using deep spatio-temporal neural networks[C]//Proceedings of the Eighteenth ACM International Symposium on Mobile Ad Hoc Networking and Computing. New York: ACM, 2018: 231-240.
[18] ZHANG C T, ZHANG H X, QIAO J P, et al. Deep transfer learning for intelligent cellular traffic prediction based on cross-domain big data[J]. IEEE Journal on Selected Areas in Communications, 2019, 37(6): 1389-1401.
[19] LIU Q Y, LI J W, LU Z M. ST-tran: spatial-temporal transformer for cellular traffic prediction[J]. IEEE Communications Letters, 2021, 25(10): 3325-3329.
[20] HU Y H, ZHOU Y J, SONG J P, et al. Citywide mobile traffic forecasting using spatial-temporal downsampling transformer neural networks[J]. IEEE Transactions on Network and Service Management, 2023, 20(1): 152-165.
[21] YAO Y, GU B, SU Z, et al. MVSTGN: a multi-view spatial-temporal graph network for cellular traffic prediction[J]. IEEE Transactions on Mobile Computing, 2023, 22(5): 2837-2849.
[22] RAO Z H, XU Y Y, PAN S M, et al. Cellular traffic prediction: a deep learning method considering dynamic nonlocal spatial correlation, self-attention, and correlation of spatiotemporal feature fusion[J]. IEEE Transactions on Network and Service Management, 2023, 20(1): 426-440.
[23] SHI X J, CHEN Z R, WANG H, et al. Convolutional LSTM network: a machine learning approach for precipitation nowcasting[C]//Proceedings of the Neural Information Processing Systems, 2015.
[24] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems, 2017.
[25] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 7132-7141.
[26] BARLACCHI G, DE NADAI M, LARCHER R, et al. A multi-source dataset of urban life in the city of Milan and the Province of Trentino[J]. Scientific Data, 2015, 2: 150055.
[27] WANG Y B, WU H X, ZHANG J J, et al. PredRNN: a recurrent neural network for spatiotemporal predictive learning[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(2): 2208-2225.