零担物流时序预测的SARIMA-GRU-BPNN组合模型及应用

doi:10.3778/j.issn.1002-8331.2309-0081

摘要/Abstract

摘要： 针对零担物流的需求物流量显著季节性、非线性和随机性特征使其预测难度大的问题，提出一种零担物流时序预测的SARIMA-GRU-BPNN组合模型的预测方法。使用季节性分解模型将物流量分解为趋势、季节性及残差，对趋势分量采用季节性差分自回归移动平均模型（SARIMA）拟合线性变化，对季节性分量采用门控循环神经网络（GRU）拟合季节性变化，对残差分量采用反向传播神经网络（BPNN）拟合非线性及随机性变化，组合重构得到最终预测值。实验结果表明，与自身单一模型SARIMA、GRU及BPNN相比，均方根误差（RMSE）分别降低31.5%、34.5%及47.1%；与其他单一模型灰色模型、支持向量机、长短期记忆网络及多元线性回归相比，RMSE分别降低71.3%、68.9%、54.4%及70.7%；与组合模型ARIMA-GRU、ARIMA-BPNN及ARIMA-SVM相比，RMSE分别降低31.0%、43.0%及56.1%，且趋势和季节性分量预测模型拟合优度达到92%和99%，有效降低整体预测误差，提升了预测精度和模型稳健性。

关键词: 零担物流, 需求预测, 时序分解, 组合模型, 人工神经网络

Abstract: Aiming at the problem that the significant seasonal, nonlinear, and stochastic characteristics of the demanded material flow of less-than-truck-load logistics (LTL) make it difficult to predict, a prediction method of SARIMA-GRU-BPNN combined model for LTL time series prediction is proposed. The seasonal decomposition model is used to decompose the logistics flow into trend, seasonality, and residual, the seasonal difference autoregressive moving average model (SARIMA) is used to fit the linear change for the trend component, the gated recurrent neural network (GRU) is used to fit the seasonal change for the seasonal component, and the back-propagation neural network (BPNN) is used to fit the nonlinear and stochastic change for the residual component, and the combination reconstruction is used to get the final prediction value. Based on the experimental results, the root mean square error (RMSE) is decreased by 31.5%, 34.5%, and 47.1% when compared to single self-contained models SARIMA, GRU, and BPNN, respectively. Additionally, the RMSE is reduced by 71.3%, 68.9%, 54.4%, and 70.7% when compared to other single models gray model, support vector machines, long and short-term memory networks, and multiple linear regression, respectively. Furthermore, the RMSE is reduced by 71.3%, 68.9%, and 54.4% when compared to combined models gray model, support vector machines, and long and short-term memory networks, respectively. In comparison to combined models ARIMA-GRU, ARIMA-BPNN, and ARIMA-SVM, the RMSE is reduced by 31.0%, 43.0%, and 56.1%, respectively. The goodness-of-fit of the trend and seasonal component prediction models reaches 92% and 99%, effectively reducing the overall prediction error and improving prediction accuracy and model robustness.

Key words: less-than-truck-load logistics (LTL), demand prediction, chronological decomposition, combinatorial model, artificial neural networks

秦音, 郭杜杜, 周飞, 王庆庆, 王洋. 零担物流时序预测的SARIMA-GRU-BPNN组合模型及应用[J]. 计算机工程与应用, 2024, 60(19): 297-308.

QIN Yin, GUO Dudu, ZHOU Fei, WANG Qingqing, WANG Yang. Combined SARIMA-GRU-BPNN Model for LTL Logistics Time Series Prediction and Application[J]. Computer Engineering and Applications, 2024, 60(19): 297-308.

参考文献

[1] 陈伟光, 聂世坤. 构建新发展格局: 基于国家治理与全球治理互动的逻辑[J]. 学术研究, 2022(1): 88-95.
CHEN W G, NIE S K. Constructing a new development pattern: based on the logic of interaction between national governance and global governance[J]. Academic Research, 2022(1): 88-95.
[2] 肖艳. 零担物流市场变革下智能物流与供应链一体化平台创新研究[J]. 商业经济研究, 2019(8): 101-103.
XIAO Y. Research on the innovation of intelligent logistics and supply chain integration platform under the change of LTL logistics market[J]. Business and Economic Research, 2019(8): 101-103.
[3] 郑永有, 杜艳. 基于灰色GM (1, 1) 模型的上海市航空物流预测研究[J]. 中国储运, 2022(10): 76-77.
ZHENG Y Y, DU Y. Research on Shanghai air logistics forecasting based on gray GM(1, 1) model[J]. China Storage and Transportation, 2022(10): 76-77.
[4] 冯杰. 基于支持向量回归机的区域物流需求预测研究[J]. 广西质量监督导报, 2021(3): 165-166.
FENG J. Research on regional logistics demand forecasting based on support vector regression machine[J]. Guangxi Quality Supervision Guide Periodical, 2021(3): 165-166.
[5] 杨耀红, 刘德福, 韩兴忠, 等. 基于LSTM-GRU模型的TBM掘进参数时序预测研究[J]. 水力发电, 2023, 49(2): 78-84.
YANG Y H, LIU D F, HAN X Z, et al. Research on time series prediction of TBM tunneling parameters based on LSTM-GRU model[J]. Water Power, 2023, 49(2): 78-84.
[6] 李芳. 基于适应性Lasso+GM(1, 1)模型的湖南物流业预测[J]. 应用数学进展, 2022, 11(1): 566-572.
LI F. Forecast of hunan logistics industry based on adaptive Lasso+GM(1,1) model[J]. Advances in Applied Mathematics, 2022, 11(1): 566-572.
[7] YAN P, ZHANG L, FENG Z, et al. Research on logistics demand forecast of port based on combined model[J]. Journal of Physics: Conference Series, 2019, 1168(3): 032116.
[8] WEN C , YANG J , ZHANG Z , et al. Prediction of the development trend of the “Internet+” logistics industry under the “Belt and Road” strategy[J]. Computational Intelligence and Neuroscience, 2022, 2022: 4630146.
[9] 杨建成. ARIMA-SVM的物流需求预测模型[J]. 现代电子技术, 2018, 41(9): 182-186.
YANG J C. Logistics demand forecasting model based on ARIMA?SVM[J]. Modern Electronics Technique, 2018, 41(9): 182-186.
[10] YIN G, PENG J. Prediction of regional logistics heat and coupling development between regional logistics and economic systems[J]. Discrete Dynamics in Nature and Society, 2021, 2021: 1-9.
[11] HUANG L, XIE G, ZHAO W, et al. Regional logistics demand forecasting: a BP neural network approach[J]. Complex & Intelligent Systems, 2021: 1-16.
[12] LU S. Research on GDP forecast analysis combining BP neural network and ARIMA model[J]. Computational Intelligence and Neuroscience, 2021: 1-10.
[13] 张国玲, 徐学红. 一种基于ARIMA-BPNN的物流需求预测模型[J]. 控制工程, 2017, 24(5): 958-962.
ZHANG G L, XU X H. A logistics demand prediction based on ARIMA-BPNN[J]. Control Engineering of China, 2017, 24(5): 958-962.
[14] 王逸文, 王维莉. 基于LSTM-RELM组合模型的电商GMV预测研究[J]. 计算机工程与应用, 2023, 59(10): 321-327.
WANG Y W， WANG W L. Research on GMV prediction of E-commerce based on LSTM-RELM combination model[J]. Computer Engineering and Applications, 2023, 59(10): 321-327.
[15] 张兴锐, 刘畅, 陈哲, 等. 基于时空图卷积网络的机场地铁短时客流预测[J]. 计算机工程与应用, 2023, 59(8): 322-330.
ZHANG X R, LIU C, CHEN Z, et al. Short-term passenger flow prediction of airport subway based on spatio-temporal graph convolutional network[J]. Computer Engineering and Applications, 2023, 59(8): 322-330.
[16] 郑长伟, 薛哲, 梁美玉, 等. 基于时空信息增强的科技论文主题趋势预测[J]. 计算机工程与应用, 2023, 59(14): 86-93.
ZHENG C W, XUE Z, LIANG M Y, et al. Topic trend prediction of scientific papers based on spatiotemporal information enhancement[J]. Computer Engineering and Applications, 2023, 59(14): 86-93.
[17] 王庆荣, 周禹潼, 朱昌锋, 等. 时空图卷积网络下的路网交通事故风险预测[J]. 计算机工程与应用, 2023, 59(13): 266-272.
WANG Q R, ZHOU Y T, ZHU C F, et al. Road network traffic accident risk prediction based on spatio-temporal graph convolution network[J]. Computer Engineering and Applications, 2023, 59(13): 266-272.
[18] CAI B, WANG Y, HUANG C, et al. GLSNN network: a multi-scale spatiotemporal prediction model for urban traffic flow[J]. Sensors, 2022, 22(22): 8880.
[19] 文家璇, 王苗, 刘济. 基于时序分解和随机森林的时间序列多步预测算法[J]. 华东理工大学学报 (自然科学版), 2023, 49（6）: 873-881.
WEN J X, WANG M, LIU J. Time series multi-step prediction algorithm based on time series decomposition and random forest[J]. Journal of East China University of Science and Technology(Natural Science Edition), 2023, 49（6）: 873-881.
[20] 赵然杭, 甘甜, 逄晓腾, 等. 基于时间序列分解的降雨数据挖掘与预测[J]. 中国农村水利水电, 2021(11): 116-122.
ZHAO R H, GAN T，PANG X T, et al. Rainfall data mining and forecasting based on time series decomposition[J]. China Rural Water and Hydropower, 2021(11): 116-122.
[21] 夏进, 王正群, 朱世明. 基于时间序列分解的交通流量预测模型[J]. 计算机应用, 2023, 43(4): 1129-1135.
XIA J, WANG Z Q, ZHU S M. Traffic flow prediction model based on time series decomposition[J]. Journal of Computer Applications, 2023, 43(4): 1129-1135.
[22] 闫勇志, 沐年国. 基于CEEMDAN-VMD-LSTM的超高频金融时间序列预测[J]. 计算机时代, 2023(5): 102-108.
YAN Y Z, MU N G. UHF financial time series predicting based on CEEMDAN-VMD-LSTM[J]. Computer Era, 2023(5): 102-108.
[23] 邓德军, 徐洪珍, 韦诗玥. E-V-ALSTM模型的股价预测 [J]. 计算机工程与应用, 2023, 59(6): 101-112.
DENG D J, XU H Z, WEI S Y. Stock price prediction based on E-V-ALSTM model[J]. Computer Engineering and Applications, 2023, 59(6): 101-112.
[24] 孙庆港, 王呈. 改进LSTM-AE算法的电梯知识库故障征兆预测[J]. 计算机工程与应用, 2023, 59(7): 311-318.
SUN Q G, WANG C. Prediction of fault symptoms in elevator knowledge base based on improved LSTM-AE algorithm[J]. Computer Engineering and Applications， 2023, 59(7): 311-318.
[25] GOZUYILMAZ S, KUNDAKCIOGLU O E. Mathematical optimization for time series decomposition[J]. OR Spectrum, 2021, 43(3): 733-758.
[26] 毛远宏, 孙琛琛, 徐鲁豫, 等. 基于深度学习的时间序列预测方法综述[J]. 微电子学与计算机, 2023, 40(4): 8-17.
MAO Y H, SUN C C, XU L Y, et al. A survey of time series forecasting methods based on deep learning[J]. Microelectronics & Computer, 2023, 40(4): 8-17.
[27] JIN B, ZENG G, LU Z, et al. Hybrid LSTM—BPNN-to-BPNN model considering multi-source information for forecasting medium-and long-term electricity peak load[J]. Energies, 2022, 15(20): 7584.
[28] ZHANG H, XIONG W, ZHANG R, et al. Prediction of gas consumption based on LSTM-BPNN hybrid model[J]. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 2022, 44(4): 10665-10680.