动态多模型指数平滑法融合的在线预测方法

doi:10.3969/j.issn.1001-506X.2020.09.17

摘要/Abstract

摘要：

针对传统预测算法实时性差、难以适应多种数据变化情况、参数无法根据数据的实时变化进行在线调整等问题,提出了一种动态多模型指数平滑法融合的在线预测算法。该算法根据历史数据的预测误差,动态调整指数平滑法的平滑系数,以及1次、2次、3次指数平滑法的融合权值,实时得到对传感器测量参数的准确预测。仿真表明,该算法在多种数据变化情况下均优于单一采用指数平滑法。

关键词: 在线预测, 指数平滑法, 动态参数, 故障预测, 模型融合

Abstract:

To solve the problems of poor real-time for the traditional prediction algorithms, poor adaptability of the multiple data variations, and the parameters can't be adjusted online with real-time change in data, a method of real-time prediction based on fused by dynamic exponential smoothing of multi-model is proposed. In this algorithm, smoothing factors of exponential smoothing, fusion weights of single, quadratic and cubic exponential smoothing are adjusted dynamically based on the prediction error of historical data, and the sensors measuring parameters are real-time predicted accurately. The simulation shows that the algorithm is better than other exponential smoothing using independently in the case of multiple data variations.

Key words: online prediction, exponential smoothing, dynamic parameter, failure prediction, model fusion

中图分类号:

TP277
E917

杨博帆, 张琳, 张搏, 宋亚飞, 丁尔启. 动态多模型指数平滑法融合的在线预测方法[J]. 系统工程与电子技术, 2020, 42(9): 2013-2021.

Bofan YANG, Lin ZHANG, Bo ZHANG, Yafei SONG, Erqi DING. Method of online prediction based on fused by dynamic exponential smoothing of multi-model[J]. Systems Engineering and Electronics, 2020, 42(9): 2013-2021.

图/表 14

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

图11

图12

表1

表2

参考文献 33

1	HOSSENI S , BARKER K , RAMIREZ-MARQUEZ J E . A review of definitions and measures of system resilience[J]. Reliability Engineering and System Safety, 2016, (145): 47- 61.
2	RIOLLI L , SAVICKI V . Information system organizational resilience[J]. Omega, 2003, 31 (3): 227- 233. doi: 10.1016/S0305-0483(03)00023-9
3	HOLLNAGEL E , WOODS D D , LEVESON N . Resilience engineering: concepts and precepts[M]. Florida: CRC Press, 2006.
4	YOUN B D, HU C, WANG P. Resilience-driven system design of complex engineered systems[C]//Proc.of the 37th Design Automation Conference, Part A and B, 2011.
5	AARON G. Creating supply chain resilience with information communication technology[D]. Texas: University of North Texas, 2018.
6	杨拥民, 钱彦岭, 李磊, 等. 装备维修保障信息化体系结构设计概论[M]. 北京: 国防工业出版社, 2012.
	YANG Y M , QIAN Y L , LI L , et al. An introduction to the information system of system structure of equipment maintenance support[M]. Beijing: National Defence Industry Press, 2012.
7	周林, 赵杰, 冯广飞. 装备故障预测与健康管理技术[M]. 北京: 国防工业出版社, 2015.
	ZHOU L , ZHAO J , FENG G F . Equipment failure prediction and health management technology[M]. Beijing: National Defence Industry Press, 2015.
8	XUE D, EL-FARRA N H. Analysis and accommodation of communication failures in event-triggered networked distributed control systems[C]//Proc.of the American Control Conference, 2017: 1815-1820.
9	张宝珍. 国外综合诊断、预测与健康管理技术的发展及应用[J]. 计算机测量与控制, 2008, 16 (5): 591- 594.
	ZHANG B Z . Development and applications of integrated diagnostics, prognostics and health management technologies of abroad[J]. Computer Measurement & Control, 2008, 16 (5): 591- 594.
10	张侃, 刘宝平, 黄栋. 精英遗传改进的非线性灰色神经网络算子与军费开支多目标组合预测应用[J]. 系统工程与电子技术, 2018, 40 (5): 1070- 1078.
	ZHANG K , LIU B P , HUANG D . Elite genetic improved nonlinear gray neural network operator and military expenditure multi-objective combination forecasting application[J]. Systems Engineering and Electronics, 2018, 40 (5): 1070- 1078.
11	RIVEST F , KOHAR R . A new timing error cost function for binary time series prediction[J]. IEEE Trans.on Neural Networks and Learning Systems, 2020, 31 (1): 174- 185. doi: 10.1109/TNNLS.2019.2900046
12	王亚男, 雷英杰, 王毅, 等. 基于直觉模糊推理的直觉模糊时间序列模型[J]. 系统工程与电子技术, 2016, 38 (6): 1332- 1338.
	WANG Y N , LEI Y J , WANG Y , et al. Intuitionistic fuzzy time series model based on intuitionistic fuzzy reasoning[J]. Systems Engineering and Electronics, 2016, 38 (6): 1332- 1338.
13	LU Y . A simple parameter-driven binary time series model[J]. Journal of Forecasting, 2020, 39 (2): 187- 199.
14	LI S Y , ZHUANG J , SHEN S F . Dynamic forecasting conditional probability of bombing attacks based on time-series and intervention analysis[J]. Risk Analysis, 2017, 37 (7): 1287- 1297. doi: 10.1111/risa.12679
15	张熙来, 赵俭辉, 蔡波. 针对PM2.5单时间序列数据的动态调整预测模型[J]. 自动化学报, 2018, 44 (10): 1790- 1798.
	ZHANG X L , ZHAO J H , CAI B . Prediction model with dynamic adjustment for single time series of PM2.5[J]. Acta Automatica Sinica, 2018, 44 (10): 1790- 1798.
16	黎锁平, 刘坤会. 动态指数平滑优化模型及其应用[J]. 系统工程学报, 2003, 18 (2): 163- 167. doi: 10.3969/j.issn.1000-5781.2003.02.012
	LI S P , LIU K H . Optimized dynamic exponential smoothing model and its applications[J]. Journal of Systems Engineering, 2003, 18 (2): 163- 167. doi: 10.3969/j.issn.1000-5781.2003.02.012
17	BARNARD G A . New methods of quality control[J]. Journal of the Royal Statistical Society, Series A (General), 1963, 126 (2): 255- 258. doi: 10.2307/2982365
18	BATES J M , GRANGER C W J . The combination of forecasts[J]. Journal of the Operational Research Society, 1969, 20 (4): 451- 468. doi: 10.1057/jors.1969.103
19	魏东涛, 张军凯, 朱倩, 等. 用组合模型预测备件消耗量[J]. 火力与指挥控制, 2016, 41 (9): 170- 172. doi: 10.3969/j.issn.1002-0640.2016.09.038
	WEI D T , ZHANG J K , ZHU Q , et al. A combined model to prediction spare parts consumption[J]. Fire Control & Command Control, 2016, 41 (9): 170- 172. doi: 10.3969/j.issn.1002-0640.2016.09.038
20	MENDES D T , LUIZ C O F . Improving time series forecasting: an approach combining bootstrap aggregation, clusters and exponential smoothing[J]. International Journal of Forecasting, 2018, 34 (4): 748- 761. doi: 10.1016/j.ijforecast.2018.05.006
21	BUI G C, ANH D T. An application of similarity search in streaming time series under DTW: online forecasting[C]//Proc.of the 8th International Symposium on Information and Communication Technology, 2017: 10-17.
22	SOARES E , COSTA P , COSTA B , et al. Ensemble of evolv-ing data clouds and fuzzy models for weather time series prediction[J]. Applied Soft Computing, 2018, 64, 445- 453. doi: 10.1016/j.asoc.2017.12.032
23	ZHANG R T, ZOU Q. Time series prediction and anomaly detection of light curve using LSTM neural network[C]//Proc.of the 2nd International Conference on Artificial Intelligence, Automation and Control Technologies, 2018, 1061: 012012.
24	RIVEST F , KOHAR R . A new timing error cost function for binary time series prediction[J]. IEEE Trans.on Neural Networks and Learning Systems, 2020, 31 (1): 174- 185. doi: 10.1109/TNNLS.2019.2900046
25	NIU T , ZHANG L , WEI S J , et al. Study on a combined prediction method based on BP neural network and improved verhulst model[J]. Systems Science & Control Engineering, 2019, 7 (3): 36- 42.
26	SEPP H , JURGEN S . Long short-term memory[J]. Neural Computation, 1997, 9 (8): 1735- 1780. doi: 10.1162/neco.1997.9.8.1735
27	Jaeger H. The "echo state" approach to analysing and training recurrent neural networks-with an erratum note[R]. Fraunhofer Institute for Autonomous Intelligent Systems, 2010.
28	ZACHARIAH C, HUMZA S, DHIREESHA K. Analysis of wide and deep echo state networks for multiscale spatiotemporal time series forecasting[C]//Proc.of the 7th Annual Neuro-inspired Computational Elements Workshop, 2019.
29	HARVEY A C . Arima models[M]. London: Palgrave Macmillan, 2018.
30	RAHMAN A , HASAN M M . Modeling and forecasting of carbon dioxide emissions in bangladesh using autoregressive integrated moving average (ARIMA) models[J]. Open Journal of Statistics, 2017, 7 (4): 560- 566. doi: 10.4236/ojs.2017.74038
31	LIU C, HOI S C H, ZHAO P, et al. Online ARIMA algorithms for time series prediction[C]//Proc.of the 30th AAAI Conference on Artificial Intelligence, 2016: 1867-1873.
32	HUA Z Q , XUE D M . ARIMA based time series forecasting model[J]. Recent Advances in Electrical & Electronic Engineering (Formerly Recent Patents on Electrical & Electronic Engineering), 2016, 9 (2): 93- 98.
33	PERERIA E N , SCARPIN C T , LUIZ A T J . Time series forecasting by using a neural ARIMA model based on wavelet decomposition[J]. Independent Journal of Management & Production, 2016, 7 (1): 252- 270.

预测步数	1次平滑	2次平滑	3次平滑	ARIMA	本文算法
1步预测	0.186 0	0.169 5	0.324 5	0.739 5	0.169 7
2步预测	0.295 7	0.265 9	0.720 5	0.748 9	0.237 0
3步预测	0.402 0	0.369 7	1.273 3	0.749 9	0.316 7
4步预测	0.502 7	0.477 7	1.983 7	0.750 3	0.399 4
5步预测	0.594 9	0.573 6	2.849 6	0.756 0	0.475 8

信号样式	1次平滑	2次平滑	3次平滑	ARIMA	本文算法
恒定电压	0.2	0.1	0.2	851.6	1.1
正弦波	0.2	0.1	0.1	969.5	0.9
矩形脉冲	0.2	0.1	0.1	698.4	0.9
锯齿波	0.2	0.1	0.1	727.5	0.9

[1]	蔺瑞管, 王华伟, 车畅畅, 倪晓梅, 熊明兰. 基于LSTM分类器的航空发动机预测性维护模型[J]. 系统工程与电子技术, 2022, 44(3): 1052-1059.
[2]	张保山, 张琳, 张搏, 鲁娜, 魏圣军. 基于故障风险标尺的复杂装备健康状态分类模型[J]. 系统工程与电子技术, 2020, 42(2): 489-496.
[3]	黄魁, 苏春. 基于灰色神经网络组合模型的故障预测[J]. 系统工程与电子技术, 2020, 42(1): 238-244.
[4]	李迎, 张璟, 刘庆, 张伟. 求解大规模多背包问题的高级人工鱼群算法[J]. 系统工程与电子技术, 2018, 40(3): 710-716.
[5]	侯晓东, 王永攀, 杨江平, 张宇. 基于状态的武器电子装备故障预测研究综述[J]. 系统工程与电子技术, 2018, 40(2): 360-367.
[6]	董磊, 阎芳, 王鹏. 基于修正高维云的故障映射模型[J]. 系统工程与电子技术, 2015, 37(2): 354-359.
[7]	杜占龙, 李小民, 席雷平, 张金中, 刘新海. 多分类概率极限学习机及其在剩余使用寿命预测中的应用[J]. 系统工程与电子技术, 2015, 37(12): 2777-2784.
[8]	王琳，张赟，彭文辉，徐波，王前程. 基于人工蜂群优化的支持向量回归预测方法[J]. 系统工程与电子技术, 2014, 36(2): 326-330.
[9]	赵建忠，李海军，叶文，董琪. 基于信息融合和改进UGM(1,1)模型的故障预测[J]. 系统工程与电子技术, 2013, 35(10): 2135-2140.
[10]	徐宇亮，孙际哲，陈西宏，王光明. 电子设备健康状态评估与故障预测方法[J]. Journal of Systems Engineering and Electronics, 2012, 34(5): 1068-1072.
[11]	范庚，马登武，邓力，吕晓峰. 基于灰色相关向量机的故障预测模型[J]. Journal of Systems Engineering and Electronics, 2012, 34(2): 424-428.
[12]	李波, 赵洁, 郭晋. 设备故障评估新指标及基于ARMA的预测系统[J]. Journal of Systems Engineering and Electronics, 2011, 33(1): 98-0101.
[13]	张磊，李行善，于劲松，万九卿. 基于关联向量机回归的故障预测算法[J]. Journal of Systems Engineering and Electronics, 2010, 32(7): 1540-1543.
[14]	王昊天，石健. 基于可用度模型的故障预测与健康管理方法[J]. Journal of Systems Engineering and Electronics, 2010, 32(12): 2584-2589.
[15]	张琪, 胡昌华, 乔玉坤, 蔡艳宁. 基于权值选优粒子滤波器的故障预测算法[J]. Journal of Systems Engineering and Electronics, 2009, 31(1): 221-224.