基于Informer神经网络的农产品物流需求预测分析——以华中地区为例

doi:10.12133/j.smartag.SA202302001

摘要/Abstract

摘要：

保障农产品物流稳定性即是保障民生问题的关键。对农产品物流需求的预测是合理规划农产品物流稳定性的重要保证。然而，农产品物流需求的预测实际较为复杂，预测过程中会受到各种因素影响。因此，为了保证对农产品物流需求预测的准确性，需要考虑多方面影响因素。本研究以农产品物流需求作为研究对象，利用Informer神经网络构建预测农产品物流需求的神经网络模型，以华中地区河南省、湖北省和湖南省为例，对三省的农产品物流需求进行预测。同时用长短时记忆（Long Short-Term Memory，LSTM）网络和Transformer神经网络对华中三省农产品物流进行需求预测，将三种模型预测结果进行对比。对比结果表明本研究构建的基于Informer神经网络模型预测测试误差平均百分比为3.39%，低于LSTM和Transformer神经网络模型的4.43%和4.35%。并且用该Informer神经网络模型对三省预测出的预测值与实际值结果较为接近，河南省2021年的预测值为4185.33，实际值为4048.1，误差为3.389%；湖北省2021年的预测值为2503.64，实际值2421.78，误差为3.380%；湖南省2021年的预测值，2933.31，实际值为2836.86，误差为3.340%。表明该模型对华中三省的农产品物流需求预测的结果较为准确。三省2023年的预测值高于2021年的预测值。因此，在2021年物流运输配套设施的基础上，必须保证物流运输效率，加强物流运输能力，以满足华中地区日益增长的物流需求。

关键词: 农产品物流, 需求预测, Informer神经网络, 长短时记忆神经网络, Transformer神经网络, 深度学习

Abstract:

Ensuring the stability of agricultural products logistics is the key to ensuring people's livelihood. The forecast of agricultural products logistics demand is an important guarantee for rational planning of agricultural products logistics stability. However, the forecasting of agricultural products logistics demand is actually complicated, and it will be affected by various factors in the forecasting process. Therefore, in order to ensure the accuracy of forecasting the logistics demand of agricultural products, many influencing factors need to be considered. In this study, the logistics demand of agricultural products is taken as the research object, relevant indicators from 2017 to 2021 were selected as characteristic independent variables and a neural network model for forecasting the logistics demand of agricultural products was constructed by using Informer neural network. Taking Henan province, Hubei province and Hunan province in Central China as examples, the logistics demands of agricultural products in the three provinces were predicted. At the same time, long short-term memory network (LSTM) and Transformer neural network were used to forecast the demand of agricultural products logistics in three provinces of Central China, and the prediction results of the three models were compared. The results showed that the average percentage of prediction test error based on Informer neural network model constructed in this study was 3.39%, which was lower than that of LSTM and Transformer neural network models of 4.43% and 4.35%. The predicted value of Informer neural network model for three provinces was close to the actual value. The predicted value of Henan province in 2021 was 4185.33, the actual value was 4048.10, and the error was 3.389%. The predicted value of Hubei province in 2021 was 2503.64, the actual value was 2421.78, and the error was 3.380%. The predicted value of Hunan province in 2021 was 2933.31, the actual value was 2836.86, and the error was 3.340%. Therefore, it showed that the model can accurately predict the demand of agricultural products logistics in three provinces of Central China, and can provide a basis for rational planning and policy making of agricultural products logistics. Finally, the model and parameters were used to predict the logistics demand of agricultural products in Henan, Hunan, and Hubei provinces in 2023, and the predicted value of Henan province in 2023 was 4217.13; Hubei province was 2521.47, and Hunan province was 2974.65, respectively. The predicted values for the three provinces in 2023 are higher than the predicted values in 2021. Therefore, based on the logistics and transportation supporting facilities in 2021, it is necessary to ensure logistics and transportation efficiency and strengthen logistics and transportation capacity, so as to meet the growing logistics demand in Central China.

Key words: agricultural products, demand forecast, Informer neural network, long short-term memory network (LSTM), Transformer, deep learning

中图分类号:

TP 393.01

左敏, 胡天宇, 董微, 张可心, 张青川. 基于Informer神经网络的农产品物流需求预测分析——以华中地区为例[J]. 智慧农业(中英文), 2023, 5(1): 34-43.

ZUO Min, HU Tianyu, DONG Wei, ZHANG Kexin, ZHANG Qingchuan. Forecast and Analysis of Agricultural Products Logistics Demand Based on Informer Neural Network: Take the Central China Aera as An Example[J]. Smart Agriculture, 2023, 5(1): 34-43.

图/表 9

表1

图1

图2

表2

图3

图4

图5

表 3

表4

参考文献 25

1	姚冠新, 杨妍, 徐静, 等. 新冠疫情背景下农产品物流需求预测——以兴化市为例[J]. 物流科技, 2021, 44(9): 91-95.
	YAO G X, YANG Y, XU J, et al. Forecasting of logistics demand for agricultural products under the background of COVID-19—A case study of Xinghua city[J]. Logistics sci-tech, 2021, 44(9): 91-95.
2	张继良. 基于组合模型的山东生鲜农产品物流需求预测[D]. 济南: 山东大学, 2021.
	ZHANG J L. Logistics demand forecast of fresh agricultural products in Shandong province based on combination model[D]. Jinan: Shandong University, 2021.
3	王晓平, 闫飞. 基于GA-BP模型的北京城镇农产品冷链物流需求预测[J]. 数学的实践与认识, 2019, 49(21): 17-27.
	WANG X P, YAN F. Prediction of cold chain logistics demand for agricultural products in Beijing based on GA-BP model[J]. Mathematics in practice and theory, 2019, 49(21): 17-27.
4	SHYLA M K, SHIVA K K, KUMAR D R. Image steganography using genetic algorithm for cover image selection and embedding[J]. Soft computing letters, 2021, 3: ID 100021.
5	皇甫红姣. 绵阳市生鲜农产品冷链物流需求预测研究[D]. 绵阳: 西南科技大学, 2021.
	HUANGFU H J. Research on cold chain logistics demand forecast of fresh agricultural products in Mianyang city[D]. Mianyang: Southwest University of Science and Technology, 2021.
6	刘文博. 辽宁省生鲜农产品冷链物流需求预测研究[J]. 全国流通经济, 2022(4): 11-13.
	LIU W B. Research on demand forecast of cold chain logistics of fresh agricultural products in Liaoning province[J]. China circulation economy, 2022(4): 11-13.
7	WANG M, LI X. Demand forecasting of agricultural cold chain logistics based on metabolic GM (1,1) model[J]. IOP conference series: Earth and environmental science, 2021, 831(1): ID 012018.
8	HUANG L, XIE G, ZHAO W, et al. Regional logistics demand forecasting: A BP neural network approach[J]. Complex & Intelligent Systems, 2021: 1-16.
9	徐晓燕, 杨慧敏, 吕修凯, 等. 基于山东省不同模型的物流需求预测比较研究[J]. 包装工程, 2022, 43(23): 207-215.
	XU X Y, YANG H M, LYU X K, et al. Comparative research on forecast of logistics demand in Shandong Province based on different models[J]. Packaging engineering, 2022, 43(23): 207-215.
10	VAN DER LAAN E, VAN DALEN J, ROHRMOSER M, et al. Demand forecasting and order planning for humanitarian logistics: An empirical assessment[J]. Journal of operations management, 2016, 45: 114-122.
11	LU B, PARK N. A study on model of regional logistics requirements prediction[J]. Journal of Korean navigation and port research, 2012, 36: 553-559.
12	BAISARIYEV M, BAKYTZHANULY A, SERIK Y, et al. Demand forecasting methods for spare parts logistics for aviation: A real-world implementation of the Bootstrap method[J]. Procedia manufacturing, 2021, 55: 500-506.
13	ZENG M L, LIU R M, GAO M, et al. Demand forecasting for rural E-commerce logistics: A gray prediction model based on weakening buffer operator[J]. 2022, 2022: ID 3395757.
14	ANSHUKA A, CHANDRA R, BUZACOTT A J V, et al. Spatio temporal hydrological extreme forecasting framework using LSTM deep learning model[J]. Stochastic environmental research and risk assessment, 2022, 36(10): 3467-3485.
15	LIU J J, YUAN X, ZENG J H, et al. Ensemble streamflow forecasting over a cascade reservoir catchment with integrated hydrometeorological modeling and machine learning[J]. Hydrology and earth system sciences, 2021, 26(2): 265-278.
16	XU C W, WANG Y Z, FU H, et al. Comprehensive analysis for long-term hydrological simulation by deep learning techniques and remote sensing[J]. Frontiers in earth science, 2022, 10: ID 875145.
17	DU S D, LI T R, YANG Y, et al. Deep air quality forecasting using hybrid deep learning framework[J/OL]. arXiv: , 2018.
18	JANARTHANAN R, PARTHEEBAN P, SOMASUNDARAM K, et al. A deep learning approach for prediction of air quality index in a metropolitan city[J]. Sustainable cities and society, 2021, 67: ID 102720.
19	PENG J H, CAO H L, ALI Z, et al. Intelligent reflecting surface-assisted interference mitigation with deep reinforcement learning for radio astronomy[J]. IEEE antennas and wireless propagation letters, 2022, 21(9): 1757-1761.
20	SEDAGHAT N, ROMANIELLO M, CARRICK J E, et al. Machines learn to infer stellar parameters just by looking at a large number of spectra[J]. Monthly notices of the Royal Astronomical Society, 2021, 501(4): 6026-6041.
21	LIU H L, LIN P, ZHENG W P, et al. A global eddy-resolving ocean forecast system in China-LICOM Forecast System (LFS)[J]. Journal of operational oceanography, 2021, 16(1): 15-27.
22	ZHU X M, ZU Z Q, REN S H, et al. Improvements in the regional South China Sea operational oceanography forecasting system (SCSOFSv2)[J]. Geoscientific model development, 2022, 15(3): 995-1015.
23	综合百科[EB/OL]. (2020-10-29)[2023-01-30]. .
24	ZHOU H Y, ZHANG S H, PENG J Q, et al. Informer: Beyond efficient transformer for long sequence time-series forecasting[J/OL]. arXiv:2012.07436 [cs.LG], 2020.
25	宋勇, 蔡志平. 大数据环境下基于信息论的入侵检测数据归一化方法[J]. 武汉大学学报(理学版), 2018, 64(2): 121-126.
	SONG Y, CAI Z P. Normalized method of intrusion detection data based on information theory in big data environment[J]. Journal of Wuhan university (natural science edition), 2018, 64(2): 121-126.

年份	省份	G1/亿元	G2/元	G3/元	G4/亿元	G5/亿元	G6/亿元	G7/千公顷	G8/万吨公里	G9/公斤	G10/万吨	G11/千公顷	G12/万吨	G13/万公里	G14/万公里	S/万吨	X/万	Y/千克
2017	河南	44824.9	20170	13730	4139.3	20940.3	19745.3	14732.53	6524.25	683.49	11271.39	16670.25	230114	0.54	26.78	3215.1	9829	327.1
	湖南	33828.1	23103	17160	2998.4	13459.8	17369.9	8321.97	3073.6	449.29	5432.21	10248.41	225551	0.47	23.97	2562.3	6633	386.3
	湖北	37235	23757	16938	3529	15713.9	17992.2	7956.14	2846.13	482.93	5374.29	9796.19	188107	0.42	26.95	2030.4	5904	343.9
2018	河南	49935.9	21964	15169	4311.1	22038.6	23586.2	14783.35	6648.91	693.9	11369.49	16776.68	259884	0.54	26.86	3231.4	9864	327.6
	湖南	36329.7	25241	18808	3084.2	13904.1	19341.4	8111.09	3022.90	439.41	5424.7	10095.08	229957	0.51	24.01	2483.5	6635	374.3
	湖北	42022	25815	19538	3548.2	17573.9	20899.9	7952.9	2839.47	480.49	5414.61	9807.57	204307	0.43	27.50	2015.3	5917	340.6
2019	河南	53717.8	23903	16332	4635.7	23035.6	26046.5	14713.98	6695.36	695.81	11447.8	16914.99	219024	0.65	26.98	3464.4	9901	349.9
	湖南	39894.1	27680	20479	3647.2	15401.7	20845.2	8122.79	2974.84	430.6	4730.02	10581.25	189740	0.56	24.06	2580.97	6640	388.7
	湖北	45429	28319	21567	3809.4	18723	22896.5	7815.89	2724.98	460.15	5338.46	10271.75	188133	0.52	28.9	2034.74	5927	343.3
2020	河南	54259.4	24810	16143	5354	22220.9	26684.5	14687.99	6825.8	688.02	11584.08	16737.73	219939	0.65	27.03	3777.58	9941	380
	湖南	41542.6	29380	20998	4240.7	15949.2	21352.7	8400.13	3015.12	453.91	4440.67	10409.77	200878	0.56	24.11	2700.53	6645	406.4
	湖北	43004.5	27881	19246	4133.2	15933.8	22937.6	7974.39	2727.43	467.35	5339.24	9753.15	160422	0.52	282.96	2205.51	5745	383.9
2021	河南	58887.4	26811	18391	5620.8	24331.6	28934.9	14705.13	6544.17	660.23	11282.39	16713.72	255551	0.65	27.16	4048.1	9883	409.6
	湖南	46063.1	31993	22798	4322.9	18126.1	23614.1	8504.26	3074.36	463.46	4701.37	10552.7	224465	0.56	24.19	2836.86	6622	428.4
	湖北	50012.9	30829	23846	4661.7	18952.9	26398.4	8109.24	2764.33	477.64	5589.54	9949.02	214762	0.52	29.69	2421.78	5830	415.4

环境类型	环境名称	参数
操作系统	Windows 10	64 bit
硬件信息	CPU	Intel（R） Core（TM） i5-8265U CPU @ 1.60 GHz （8 CPUs）~1.8 GHz
	GPU	Radeon 540X Series
	RAM	16 GB
软件工具	Python 3.7	Numpy 1.21.6
		Scikit_Learn 1.0.2
		Pandas 1.3.5
		Torch 1.12.1
		Matplotlib 3.5.3

折叠数量	训练错误		测试错误
折叠数量	绝对误差	误差百分比/%	绝对误差	误差百分比/%
Informer 1倍	77.21	4.61	85.96	5.17
Informer 2倍	21.49	2.17	31.19	2.34
Informer 3倍	20.87	2.23	33.47	2.53
Informer 4倍	32.94	3.54	37.26	3.59
Informer 5倍	30.37	3.43	36.61	3.36
Informer平均值	36.58	3.07	44.89	3.39
LSTM 1倍	85.37	6.31	81.96	5.97
LSTM 2倍	31.23	4.21	38.23	3.64
LSTM 3倍	22.87	3.03	35.57	3.83
LSTM 4倍	34.74	3.84	28.31	2.89
LSTM 5倍	28.37	3.75	33.18	3.16
LSTM平均值	40.52	4.37	43.45	4.43
Transformer 1倍	81.07	4.61	82.15	5.17
Transformer 2倍	39.61	3.77	47.59	2.34
Transformer 3倍	27.53	2.03	38.26	3.47
Transformer 4倍	37.81	3.54	35.81	3.19
Transformer 5倍	33.41	3.43	31.73	3.06
Transformer平均值	39.61	3.07	47.11	4.35

序号	G1	G2	G3	G4	G5	G6	G7	G8	G9	G10	G11	G12	G13	G14	X	Y	绝对误差/%
1	√	√	√	√	√	√	√	√	√	√	√	√	√	√	√	√	3.39
2	×	√	√	√	√	√	√	√	√	√	√	√	√	√	√	√	3.52
3	√	×	√	√	√	√	√	√	√	√	√	√	√	√	√	√	3.58
4	√	√	×	√	√	√	√	√	√	√	√	√	√	√	√	√	3.81
5	√	√	√	×	√	√	√	√	√	√	√	√	√	√	√	√	3.77
6	√	√	√	√	×	√	√	√	√	√	√	√	√	√	√	√	3.86
7	√	√	√	√	√	×	√	√	√	√	√	√	√	√	√	√	3.46
8	√	√	√	√	√	√	×	√	√	√	√	√	√	√	√	√	4.07
9	√	√	√	√	√	√	√	×	√	√	√	√	√	√	√	√	3.44
10	√	√	√	√	√	√	√	√	×	√	√	√	√	√	√	√	3.47
11	√	√	√	√	√	√	√	√	√	×	√	√	√	√	√	√	3.62
12	√	√	√	√	√	√	√	√	√	√	×	√	√	√	√	√	3.53
13	√	√	√	√	√	√	√	√	√	√	√	×	√	√	√	√	3.74
14	√	√	√	√	√	√	√	√	√	√	√	√	×	√	√	√	3.48
15	√	√	√	√	√	√	√	√	√	√	√	√	√	×	√	√	3.51
16	√	√	√	√	√	√	√	√	√	√	√	√	√	√	×	√	3.43
17	√	√	√	√	√	√	√	√	√	√	√	√	√	√	√	×	3.52

[1]	李豪, 杜雨秋, 肖星竹, 陈彦羲. 基于深度学习的四川盆地丘陵区县域耕地遥感识别研究[J]. 智慧农业(中英文), 2024, 6(3): 34-45.
[2]	聂刚刚, 饶洪辉, 李泽锋, 刘木华. 基于改进YOLACT的油茶叶片炭疽病感染严重程度分级模型[J]. 智慧农业(中英文), 2024, 6(3): 138-147.
[3]	张京, 赵泽瑄, 赵艳茹, 卜泓超, 吴星宇. 基于Bi-GRU和空-谱信息融合的油菜菌核病侵染区域高光谱图像分割方法[J]. 智慧农业(中英文), 2024, 6(2): 40-48.
[4]	庞春晖, 陈鹏, 夏懿, 章军, 王兵, 邹岩, 陈天娇, 康辰瑞, 梁栋. 用于小麦多生长阶段倒伏边界精准检测的分层交互特征金字塔网络[J]. 智慧农业(中英文), 2024, 6(2): 128-139.
[5]	张玉玉, 邴树营, 纪元浩, 严蓓蓓, 许金普. 基于改进YOLOv8s的玫瑰鲜切花分级方法[J]. 智慧农业(中英文), 2024, 6(2): 118-127.
[6]	张建华, 姚琼, 周国民, 吴雯迪, 修晓杰, 王健. 作物农艺性状与形态结构表型智能识别技术综述[J]. 智慧农业(中英文), 2024, 6(2): 14-27.
[7]	郭旺, 杨雨森, 吴华瑞, 朱华吉, 缪祎晟, 顾静秋. 农业大模型：关键技术、应用分析与发展方向[J]. 智慧农业(中英文), 2024, 6(2): 1-13.
[8]	王鹤榕, 陈英义, 柴莹倩, 徐玲, 于辉辉. 融合VoVNetv2和置换注意力机制的鱼群摄食图像分割方法[J]. 智慧农业(中英文), 2023, 5(4): 137-149.
[9]	李政凯, 于嘉辉, 潘时佳, 贾泽丰, 牛子杰. 冬季猕猴桃树单木骨架提取与冠层生长预测方法[J]. 智慧农业(中英文), 2023, 5(4): 92-104.
[10]	唐辉, 王铭, 于秋实, 张佳茜, 刘连涛, 王楠. 融合改进UNet和迁移学习的棉花根系图像分割方法[J]. 智慧农业(中英文), 2023, 5(3): 96-109.
[11]	泮玮婷, 孙梦丽, 员琰, 刘平. 基于深度学习ImCascade R-CNN的小麦籽粒表形鉴定方法[J]. 智慧农业(中英文), 2023, 5(3): 110-120.
[12]	管博伦, 张立平, 朱静波, 李闰枚, 孔娟娟, 汪焱, 董伟. 农业病虫害图像数据集构建关键问题及评价方法综述[J]. 智慧农业(中英文), 2023, 5(3): 17-34.
[13]	龙佳宁, 张昭, 刘晓航, 李云霞, 芮照钰, 余江帆, 张漫, FLORES Paulo, 韩哲雄, 胡灿, 王旭峰. 利用改进EfficientNetV2和无人机图像检测小麦倒伏类型[J]. 智慧农业(中英文), 2023, 5(3): 62-74.
[14]	刘易雪, 宋育阳, 崔萍, 房玉林, 苏宝峰. 基于无人机遥感和深度学习的葡萄卷叶病感染程度诊断方法[J]. 智慧农业(中英文), 2023, 5(3): 49-61.
[15]	毛克彪, 张晨阳, 施建成, 王旭明, 郭中华, 李春树, 董立新, 吴门新, 孙瑞静, 武胜利, 姬大彬, 蒋玲梅, 赵天杰, 邱玉宝, 杜永明, 徐同仁. 基于人工智能的地球物理参数反演范式理论及判定条件[J]. 智慧农业(中英文), 2023, 5(2): 161-171.