基于灰色-神经网络的民机需求组合预测

doi:10.12305/j.issn.1001-506X.2024.05.19

摘要/Abstract

摘要：

民机数量是反映民航运输能力的重要标志, 而对民机数量进行预测, 能够研究分析未来民航业的发展趋势。本文重点研究了民机需求预测的模型架构和实施方法, 首先以2013年到2020年民机数量和其他关键因素作为原始样本, 然后把2021年的数据作为检验样本, 最后通过构建灰色-神经网络组合预测模型对未来的民机需求进行预测。从预测结果来看, 灰色模型GM(1, 1)与反向传播(back propagation, BP)神经网络模型结合效果较好, 组合模型预测精度高, 充分证明了该模型的有效性和可行性, 同时预测结果对分析未来航空运输情况也具有一定的参考意义。

关键词: 民机, 神经网络, 组合预测

Abstract:

The number of civil aircraft is an important symbol that reflects the transport capacity of civil aviation. By predicting the number of civil aircraft, the development trend of civil aviation industry in the future can be studied and analyed. This paper focuses on the model architecture and implementation methods of civil aircraft demand forecasting. Firstly, the number of civil aircraft and other key factors from 2013 to 2020 are taken as the original samples, then the data of 2021 is taken as the test samples. Finally, the future demand of civil aircraft is predicted by constructing the combined prediction model of gray-neural network. From the prediction results, the combination of grey model GM (1, 1) and back propagation (BP) neural network model has good effect, and the combination model has high prediction accuracy, which fully proves the validity and feasibility of this model. Meanwhile, the prediction results will also have some reference significance for analyzing the future air transportation situation.

Key words: civil aircraft, neural network, combination prediction

中图分类号:

N945

庆豪, 方志耕, 王育红, 邱玺睿. 基于灰色-神经网络的民机需求组合预测[J]. 系统工程与电子技术, 2024, 46(5): 1665-1672.

Hao QING, Zhigeng FANG, Yuhong WANG, Xirui QIU. Combination prediction of civil aircraft demand based on grey-neural network[J]. Systems Engineering and Electronics, 2024, 46(5): 1665-1672.

图/表 17

表1

表2

图1

图2

表3

图3

表4

图4

图5

表5

图6

图7

表6

图8

图9

表7

表8

参考文献 30

1	CHEN F , XU B C , FAN D H . Using the system dynamics model on sustainable safety development of civil aviation[J]. International Journal of Technology, Policy and Management, 2022, 22 (1/2): 3- 23. doi: 10.1504/IJTPM.2022.122532
2	刘光才, 赖汪湾. 基于粗糙集理论及灰色关联分析模型的我国民航发展水平综合评价[J]. 数学的实践与认识, 2017, 47 (22): 46- 57.
	LIU G C , LAI W W . Comprehensive evaluation of civil aviation development in China based on rough set theory and grey correlation analysis model[J]. Mathematical Practice and Understanding, 2017, 47 (22): 46- 57.
3	RAJAN B P T , MUTHUKUMARAN N . Grey neural network channel estimation and RBFNN hybrid precoding schemes for the multi user millimeter wave massive MIMO[J]. Transactions on Emerging Telecommunications Technologies, 2022, 34 (2): 1733- 1755.
4	陈雄寅. 基于BP神经网络的港口物流需求预测研究[J]. 物流工程与管理, 2022, 44 (12): 11-14, 20. doi: 10.3969/j.issn.1674-4993.2022.12.003
	CHEN X Y . Research on port logistics demand forecast based on BP neural network[J]. Logistics Engineering and Management, 2022, 44 (12): 11-14, 20. doi: 10.3969/j.issn.1674-4993.2022.12.003
5	ZHANG L , XUE H B , LI Z Y , et al. Robust stability analysis of switched grey neural network models with distributed delays over C[J]. Grey Systems: Theory and Application, 2022, 12 (4): 879- 896. doi: 10.1108/GS-11-2021-0177
6	黄秋红, 王霄, 杨靖, 等. 基于集群划分的区域短期风电功率预测模型[J]. 电力科学与工程, 2022, 38 (12): 8- 17. doi: 10.3969/j.ISSN.1672-0792.2022.12.002
	HUNAG Q H , WANG X , YANG J , et al. Regional short-term wind power prediction model based on cluster division[J]. Electric Power Science and Engineering, 2022, 38 (12): 8- 17. doi: 10.3969/j.ISSN.1672-0792.2022.12.002
7	唐贵基, 周威, 王晓龙, 等. 基于变维GRU-BiLSTM神经网络模型的滚动轴承寿命预测[J]. 中国工程机械学报, 2022, 20 (6): 498- 503.
	TANG G J , ZHOU W , WANG X L , et al. Rolling bearing life prediction based on variable dimension GRU-BiLSTM neural network model[J]. Chinese Journal of Construction Machinery, 2022, 20 (6): 498- 503.
8	陈凯杰, 唐振鹏, 吴俊传, 等. 贵金属期货价格预测方法及实证研究[J]. 中国管理科学, 2022, 30 (12): 245- 253.
	CHEN K J , TANG Z P , WU J C , et al. Prediction method and empirical study of precious metal futures price[J]. China Management Science, 2022, 30 (12): 245- 253.
9	SHENG Y F , ZHANG J J , TAN W , et al. Application of grey model and neural network in financial revenue forecast[J]. Computers, Materials & Continua, 2021, 69 (3): 4043- 4059.
10	UTSUMI M , KITADA K , TOKUNAGA N , et al. A combined prediction model for biliary tract cancer using the prognostic nutritional index and pathological findings: a single-center retrospective study[J]. BMC Gastroenterology, 2021, 21 (1) doi: 10.1186/s12876-021-01957-5
11	KUCHANSKY A , BILOSHCHYTSKYI A , ANDRASHKO Y , et al. Development of adaptive combined models for predicting time series based on similarity identification[J]. Eastern-European Journal of Enterprise Technologies, 2018, 1 (4): 32- 42.
12	MALEK S, ABDELLATIF E A. Intelligent system-based support vector regression for supply chain demand forecasting[C]// Proc. of the 2nd World Conference on Complex Systems, 2014: 79-83.
13	鲁玉芬, 方从严, 开明. 灰色-时序组合模型在建筑物沉降预测中的应用[J]. 科技和产业, 2022, 22 (8): 315- 318. doi: 10.3969/j.issn.1671-1807.2022.08.049
	LU Y F , FANG C Y , KAI M . Application of gray-time series combination model in building settlement prediction[J]. Science and Technology Industry, 2022, 22 (8): 315- 318. doi: 10.3969/j.issn.1671-1807.2022.08.049
14	郭为民. 灰色理论和加权平均法组合预测家用空调产品销售量[J]. 内蒙古科技与经济, 2021, (15): 61- 62.
	GUO W M . Forecasting sales volume of household air conditioning products by combination of grey theory and weighted average method[J]. Inner Mongolia Science and Technology and Economy, 2021, (15): 61- 62.
15	邵梦汝, 程天伦, 马晓晨. 基于灰色神经网络的铁路货运量组合预测[J]. 交通运输工程与信息学报, 2016, 14 (3): 129- 135. doi: 10.3969/j.issn.1672-4747.2016.03.019
	SHAO M R , CHEN T L , MA X C . Railway freight volume combination forecasting based on grey neural network[J]. Journal of Transportation Engineering and Information, 2016, 14 (3): 129- 135. doi: 10.3969/j.issn.1672-4747.2016.03.019
16	陈莎莎. 基于组合预测模型下某区碳排放的预测研究[J]. 中国新技术新产品, 2022, (12): 130- 132.
	CHEN S S . Based on the combination forecast model district of carbon emissions prediction research[J]. China's New Technology and New Products, 2022, (12): 130- 132.
17	SUN Z H , FAN Y Q . A combined water quality pollution prediction model based on the spark big data platform[J]. AQUA—Water Infrastructure, Ecosystems and Society, 2022, 71 (9): 963- 974.
18	SHARMA R C , DABRA V , SINGH G , et al. Multi response optimization while machining of stainless steel 316L using intelligent approach of grey theory and grey-TLBO[J]. World Journal of Engineering, 2022, 19 (3): 329- 339. doi: 10.1108/WJE-06-2020-0226
19	LIU Z Y , ZHANG X K , DONG Z B . TSF-transformer: a time series forecasting model for exhaust gas emission using transformer[J]. Applied intelligence (Dordrecht, Netherlands), 2022, 53 (13): 11- 15.
20	HU Y T , LI S C , LI S Y , et al. Research on the combined prediction model of milling sound pressure level based on force-thermal-vibration multi-feature fusion[J]. The International Journal of Advanced Manufacturing Technology, 2021, 115 (1/2): 233- 245.
21	SUN X X , WANG Y Q , MENG W J . Evaluation system of curved conveyor belt deviation state based on the ARIMA- LSTM combined prediction model[J]. Machines, 2022, 10 (11): 1042- 1043.
22	XIE N M . A summary of grey forecasting models[J]. Grey Systems: Theory and Application, 2022, 12 (4): 703- 722.
23	HU Y C . Forecasting the demand for tourism using combinations of forecasts by neural network-based interval grey prediction models[J]. Asia Pacific Journal of Tourism Research, 2021, 26 (12): 1350- 1363.
24	ZHU J N , LIU X C , LIU C . Non-equidistant non-homogenous grey prediction model with fractional accumulation and its application[J]. Journal of Intelligent & Fuzzy Systems, 2021, 40 (6): 11861- 11874.
25	MIZUNO N , KUBOSHIMA R . Implementation and evaluation of non-linear optimal feedback control for ship' s automatic berthing by recurrent neural network[J]. IFAC Papers Online, 2019, 52 (21): 91- 96.
26	LING X , LIU S P , LIU Q , et al. Research on remaining service life prediction of platform screen doors system based on genetic algorithm to optimize BP neural network[J]. Enterprise Information Systems, 2022, 16 (8/9): 198- 204.
27	DAI C , ZHUANG G W , WANG J , et al. Error correction method of transformer harmonic detection based on the improved BP neural network[J]. Journal of Physics: Conference Series, 2023, 2419 (1): 572- 579.
28	JIYEON K , YULIM S , EUNJUNG C . An intrusion detection model based on a convolutional neural network[J]. Journal of Multimedia Information System, 2019, 6 (4): 165- 172.
29	HE Y Y. Feature analysis of mechanical and electronic signals based on grey neural network prediction[C]//Proc. of the 4th International Conference on Education, Management and Information Technology, 2018: 1431-1436.
30	LI Q H , MA C L , WANG C Y , et al. Application of combined prediction model in surface roughness prediction[J]. Journal of Nanoelectronics and Optoelectronics, 2022, 17 (11): 1511- 1516.

年份	X₁	X₂	X₃	X₄	X₅	Y
2013	9.53	671.72	35 397.00	592 963.00	2 876.00	2 145.00
2014	9.51	748.12	39 195.00	635 910.00	3 142.00	2 370.00
2015	9.49	851.65	43 618.00	688 858.20	3 326.00	2 650.00
2016	9.41	962.51	48 796.00	746 395.10	3 794.00	2 950.00
2017	9.49	1 083.08	55 156.00	832 035.90	4 418.00	3 296.00
2018	9.36	1 206.53	61 173.77	919 281.10	4 945.00	3 639.00
2019	9.33	1 293.25	65 993.42	986 515.20	5 521.00	3 818.00
2020	6.49	798.51	41 777.82	1 015 986.20	5 581.00	3 903.00
2021	6.62	856.75	44 055.74	1 143 670.00	4 864.00	3 946.00

年份	原始值	拟合值	误差	相对误差/%
2013	2 145	2 145	0	0.0
2014	2 370	2 492	-122	-5.1
2015	2 650	2 706	-56	-2.1
2016	2 950	2 938	12	0.4
2017	3 296	3 190	106	3.2
2018	3 639	3 464	175	4.8
2019	3 818	3 761	57	1.5
2020	3 903	4 083	-180	-4.6
2021	3 946	4 433	-487	-12.4

年份	X₁	X₂	X₃	X₄	X₅	Y
2013	0.120 3	0.079 3	0.081 3	0.078 4	0.074 8	0.074 7
2014	0.120 0	0.088 3	0.090 1	0.084 1	0.081 7	0.082 5
2015	0.119 8	0.100 5	0.100 2	0.091 1	0.086 5	0.092 3
2016	0.118 8	0.113 6	0.112 1	0.098 7	0.098 6	0.102 7
2017	0.119 8	0.127 8	0.126 7	0.110 0	0.114 9	0.114 8
2018	0.118 1	0.142 4	0.140 6	0.121 6	0.128 6	0.126 7
2019	0.117 8	0.152 6	0.151 7	0.130 5	0.143 5	0.133 0
2020	0.081 9	0.094 3	0.096 0	0.134 4	0.145 1	0.135 9
2021	0.083 6	0.101 1	0.101 2	0.151 2	0.126 4	0.137 4

年份	原始值	拟合值	误差	相对误差/%
2013	2 145	2 188	-43	-2.0
2014	2 370	2 322	48	2.0
2015	2 650	2 588	62	2.3
2016	2 950	2 966	-16	-0.5
2017	3 296	3 223	73	2.2
2018	3 639	3 682	-43	-1.2
2019	3 818	3 713	105	2.8
2020	3 903	3 948	-45	-1.2
2021	3 946	3 572	374	9.5

年份	X₁	X₂	X₃	X₄	X₅	X₆	Y
2013	0.120 3	0.079 3	0.081 3	0.078 4	0.074 8	0.073 4	0.074 7
2014	0.120 0	0.088 3	0.090 1	0.084 1	0.081 7	0.085 3	0.082 5
2015	0.119 8	0.100 5	0.100 2	0.091 1	0.086 5	0.092 6	0.092 3
2016	0.118 8	0.113 6	0.112 1	0.098 7	0.098 6	0.100 6	0.102 7
2017	0.119 8	0.127 8	0.126 7	0.110 0	0.114 9	0.109 2	0.114 8
2018	0.118 1	0.142 4	0.140 6	0.121 6	0.128 6	0.118 6	0.126 7
2019	0.117 8	0.152 6	0.151 7	0.130 5	0.143 5	0.128 7	0.133 0
2020	0.081 9	0.094 3	0.09 6	0.134 4	0.145 1	0.139 8	0.135 9
2021	0.083 6	0.101 1	0.101 2	0.151 2	0.126 4	0.151 8	0.137 4