销量约束下基于切片递归神经网络模型的成品油价格推荐算法

doi:10.7523/j.ucas.2021.0056

摘要/Abstract

摘要： 加油站成品油零售价格的确定是智慧加油站发展的关键。由于成品油价格的变化遵循复杂的非线性规律，尽管以长短期记忆（LSTM）为代表的非线性时序模型提高了传统时序预测方法的精度，但其运行效率难以满足动态变化的油价预测需求。针对这一问题，提出一种基于切片递归神经网络（SRNN）的成品油价格推荐模型，该模型以LSTM模型为递归单元，创新性地通过决策者根据多源数据得到的聚类结果筛选、设置的市场环境因子，对成品油销量施加影响，从而实现在销售约束条件下的成品油价格推荐。基于4年的加油站历史数据对模型预测性能进行了评估。结果表明，使用该模型与LSTM神经网络具有相同的预测精度水平，但比LSTM神经网络的运行速度快72倍。此外，基于SRNN模型的成品油价格推荐算法，加油站在实际销售中得到有效的应用，验证该模型的实用价值。

关键词: 长短期记忆人工神经网络, 价格推荐算法, 智慧加油站, 条件切片循环人工神经网络, 不完整多视角聚类

Abstract: Determining the retail fuel price for the petrol stations is essential for the development of smart petrol stations. Since the changes in the fuel price follow a complex nonlinear model, the nonlinear time series mode represented by long short-term memory (LSTM) have improved the accuracy of traditional time series forecasting methods, though, its running efficiency is still difficult to meet the dynamic demand of oil price forecasting. To address this issue, in this paper we propose a fuel price recommendation model based on the sliced recurrent neural network (SRNN) with an LSTM model as the recurrent unit under the sales constraints. We further train this model and evaluate its learning parameters, such as learning rate, based on 4 years of historical data. In our evaluations, we utilize real data from petrol stations. Results show that our proposed model achieves the same level of accuracy as that of the LSTM neural network; however, it is 72 times faster than that of the LSTM neural network. Besides, the fuel price recommendation model based on the SRNN is efficiently applied to real petrol stations hence confirmed its practical value.

Key words: long short-term memory network (LSTM), price recommendation, smart gas station, conditional sliced recurrent neural network, incomplete multi-view clustering

中图分类号:

TP399

连会强, 刘兵, 李朋远, 于华. 销量约束下基于切片递归神经网络模型的成品油价格推荐算法[J]. 中国科学院大学学报, 2023, 40(4): 566-576.

LIAN Huiqiang, LIU Bing, LI Pengyuan, YU Hua. A fuel price recommendation model based on the sliced recurrent neural network under sales constraints[J]. Journal of University of Chinese Academy of Sciences, 2023, 40(4): 566-576.

参考文献

[1] Klein T, Walther T. Oil price volatility forecast with mixture memory GARCH[J]. Energy Economics, 2016, 58:46-58.DOI:10.1016/j.eneco.2016.06.004.
[2] Baumeister C, Guérin P, Kilian L. Do high-frequency financial data help forecast oil prices? The MIDAS touch at work[J]. International Journal of Forecasting, 2015, 31(2):238-252.DOI:10.1016/j.ijforecast.2014.06.005.
[3] Xiang Y, Zhuang X H. Application of ARIMA model in short-term prediction of international crude oil price[J]. Advanced Materials Research, 2013, 798/799:979-982.DOI:10.4028/www.scientific.net/amr.798-799.979.
[4] Hu J W S, Hu Y C, Lin R R W. Applying neural networks to prices prediction of crude oil futures[J]. Mathematical Problems in Engineering, 2012, 2012:1-12.DOI:10.1155/2012/959040.
[5] Kristjanpoller W, Minutolo M C. Forecasting volatility of oil price using an artificial neural network-GARCH model[J]. Expert Systems With Applications, 2016, 65:233-241.DOI:10.1016/j.eswa.2016.08.045.
[6] Elman J L. Finding structure in time[J]. Cognitive Science, 1990, 14(2):179-211.DOI:10.1016/0364-0213(90)90002-E.
[7] Jordan M I. Serial order:a parallel distributed processing approach[J]. Advances in Psychology, 1997, 121:471-495.DOI:10.1016/50166-4115(97)80111-2.
[8] 何树红, 杨博, 戴明爽. 基于动态递归神经网络的石油价格预测[J]. 云南民族大学学报(自然科学版), 2013, 22(1):31-35.DOI:10.3969/j.issn.1672-8513.2013.01.008.
[9] 赵曦. 投影寻踪和神经网络算法的石油价格预测[J]. 计算机仿真, 2012, 29(5):371-374. DOI:10.3969/j.issn.1006-9348.2012.05.091.
[10] Bengio Y, Simard P, Frasconi P. Learning long-term dependencies with gradient descent is difficult[J]. IEEE Transactions on Neural Networks, 1994, 5(2):157-166.DOI:10.1109/72.279181.
[11] Chaitanya Lahari M, Ravi D H, Bharathi R. Fuel price prediction using RNN[C]//2018 International Conference on Advances in Computing, Communications and Informatics (ICACCI). September 19-22, 2018, Bangalore, India. IEEE, 2018:1510-1514.DOI:10.1109/ICACCI.2018.8554642.
[12] Wu Y X, Wu Q B, Zhu J Q. Improved EEMD-based crude oil price forecasting using LSTM networks[J]. Physica A:Statistical Mechanics and Its Applications, 2019, 516:114-124.DOI:10.1016/j.physa.2018.09.120.
[13] Chen Y H, He K J, Tso G K F. Forecasting crude oil prices:a deep learning based model[J]. Procedia Computer Science, 2017, 122:300-307.DOI:10.1016/j.procs.2017.11.373.
[14] Cen Z P, Wang J. Crude oil price prediction model with long short term memory deep learning based on prior knowledge data transfer[J]. Energy, 2019, 169:160-171.DOI:10.1016/j.energy.2018.12.016.
[15] Yao T X, Wang Z H. Crude oil price prediction based on LSTM network and GM (1, 1) model[J]. Grey Systems:Theory and Application, 2020, 11(1):80-94.DOI:10.1108/gs-03-2020-0031.
[16] Bristone M, Prasad R, Abubakar A A. CPPCNDL:crude oil price prediction using complex network and deep learning algorithms[J]. Petroleum, 2020, 6(4):353-361.DOI:10.1016/j.petlm.2019.11.009.
[17] Nagendra Kumar Y J, Preetham P, Kiran Varma P, et al. Crude oil price prediction using deep learning[C]//2020 Second International Conference on Inventive Research in Computing Applications (ICIRCA). July 15-17, 2020, Coimbatore, India. IEEE, 2020:118-123.DOI:10.1109/ICIRCA48905.2020.9183258.
[18] Li Z K, Wang M, Wang X H, et al. Oil price forecasting based on variational mode decomposition, relative entropy and LSTM neural network[J]. IOP Conference Series:Materials Science and Engineering, 2020, 750:012203.DOI:10.1088/1757-899X/750/1/012203.
[19] Orojo O, Tepper J, McGinnity T M, et al. A multi-recurrent network for crude oil price prediction[C]//2019 IEEE Symposium Series on Computational Intelligence (SSCI). December 6-9, 2019, Xiamen, China. IEEE, 2020:2940-2945.DOI:10.1109/SSCI44817.2019.9002841.
[20] Biswas S, Gall J. Structural recurrent neural network (SRNN) for group activity analysis[C]//2018 IEEE Winter Conference on Applications of Computer Vision (WACV). March 12-15, 2018, Lake Tahoe, NV, USA. IEEE, 2018:1625-1632.DOI:10.1109/WACV.2018.00180.
[21] 詹静,范雪,刘一帆,等. SEMBeF:一种基于分片循环神经网络的敏感高效的恶意代码行为检测框架[J]. 信息安全学报, 2019, 4(6):67-79.DOI:10.19363/J.cnki.cn10-1380/tn.2019.11.06.
[22] Liu S Y, Hao X G, Meng Z J, et al. Application of SRNN-GRU in photovoltaic power forecasting[J]. E3S Web of Conferences, 2021, 256:02001.DOI:10.1051/e3sconf/202125602v01.
[23] Li B, Cheng Z H, Xu Z H, et al. Long text analysis using sliced recurrent neural networks with breaking point information enrichment[C]//ICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). May 12-17, 2019, Brighton, UK. IEEE, 2019:7550-7554.DOI:10.1109/ICASSP.2019.8683812.
[24] Pei X J, Tian S W, Yu L, et al. A two-stream network based on capsule networks and sliced recurrent neural networks for DGA botnet detection[J]. Journal of Network and Systems Management, 2020, 28(4):1694-1721.DOI:10.1007/S10922-020-09554-9.
[25] Yu Z P, Shi L Q, Liu G S. Dissect sliced-RNN in multi-attention view[J]. Australian Journal of Intelligent Information Processing System, 2019, 17(2):61-68.
[26] Zhang L, Wu N, Ge F, et al. A dynamic branch predictor based on parallel structure of SRNN[J]. IEEE Access, 2020, 8:86230-86237.DOI:10.1109/ACCESS.2020.2992643.
[27] Hochreiter S, Schmidhuber J. Long short-term memory[J]. Neural Computation, 1997, 9(8):1735-1780.DOI:10.1162/neco.1997.9.8.1735.
[28] Kolen J F, Kremer S C. A field guide to dynamical recurrent networks[M]. New York:IEEE Press, 2001.
[29] Sherstinsky A. Fundamentals of recurrent neural network (RNN) and long short-term memory (LSTM) network[J]. Physica D:Nonlinear Phenomena, 2020, 404:132306.DOI:10.1016/j.physd.2019.132306.
[30] Lian H Q, Xu H Y, Wang S W, et al. Partial multiview clustering with locality graph regularization[J]. International Journal of Intelligent Systems, 2021, 36(6):2991-3010.DOI:10.1002/int.22409.
[31] Moyo V, Sibanda K. The generalization ability of artificial neural networks in forecasting TCP/IP traffic trends:how much does the size of learning rate matter?[J]. International Journal of Computer Science and Application, 2015, 4(1):9-17. DOI:10.12783/ijcsa.2015.0401.02.