基于非线性混沌时序动力系统的预测方法研究

马军海; 陈予恕; 辛宝贵

基于非线性混沌时序动力系统的预测方法研究

1.
天津大学管理学院,天津 300072;2.天津大学力学系,天津 300072;3.天津财经大学,天津 300222

基金项目: 国家自然科学基金资助项目(70271071,9990510)

详细信息

作者简介:
马军海(1965- ),男,山东青岛人,教授,博士,博士生导师,已在国内外核心期刊发表论文30余篇,主要研究方向:复杂非线性动力系统、复杂混沌时序重构及其工程应用(联系人.Tel:+86-22-81593670;E-mail:lzqsly@126.com);陈予恕(1931- ),男,山东肥城人,教授、博士生指导教师,俄罗斯应用科学院外籍院士.

中图分类号: O175.14；O241.81
计量
- 文章访问数: 2814
- HTML全文浏览量: 122
- PDF下载量: 644
- 被引次数: 0
出版历程
- 收稿日期: 2002-04-09
- 修回日期: 2003-12-23
- 刊出日期: 2004-06-15

Study on Prediction Methods for Dynamic Systems of Nonlinear Chaotic Time Series

1.
Management School, Tianjin University, Tianjin 300072, P. R. China;

摘要

摘要: 主要研究由混沌时序所确定的非线性动力系统的预测方法．研究了非线性自相关混沌模型的结构,模型阶数的确立技术．将神经网络和小波理论相结合，研究了小波变换神经网络的结构，给出了小波神经网络的学习方法；提出了一种新的基于小波网络的参数辨识方法．该方法可以有选择地提取时序中的不同的时间、频率尺度,实现原时序的趋势或细节预测．通过对混沌时序进行预处理，并比较预处理后的预测结果,得到了一些有益的结果：用非线性自相关混沌模型采用小波网络对模型参数进行辨识,其辨识的准确程度较高,用该模型对混沌时序(包括含有噪声)的预测比较有效．
- 非线性自相关混沌模型 /
- 小波神经网络 /
- 参数识别 /
- 时序预测
Abstract: The prediction methods for nonlinear dynamic systems which are decided by chaotic time series are mainly studied as well as structures of nonlinear self-related chaotic models and their dimensions. By combining neural networks and wavelet theories, the structures of wavelet transform neural networks were studied and also a wavelet neural networks learning method was given. Based on wavelet networks, a new method for parameter identification was suggested, which can be used selectively to extract different scales of frequency and time in time series in order to realize prediction of tendencies or details of original time series. Through pre-treatment and comparison of results before and after the treatment, several useful conclusions are reached: High accurate identification can be guaranteed by applying wavelet networks to identify parameters of self-related chaotic models and more valid prediction of the chaotic time series including noise can be achieved accordingly.
- nonlinear self-related chaotic model /
- wavelet neural network /
- parameter identification /
- time series prediction

HTML全文

参考文献(13)

[1]	LIANG Yue-cao,HOGN Yi-guang,FANG Hai-ping,et al.Predicting chaotic time series with wavelet networks[J].Phys D,1995,85(8):225—238. doi: 10.1016/0167-2789(95)00119-O
[2]	ZHANG Qing-hua. Wavelet Networks[J].IEE Transactions on Neural Networks,1992,11(6):889—898.
[3]	Castillo E,Gutierrez J M. Nonlinear time series modeling and prediction using functional networks extracting information masked by chaos[J].Phys Lett A,1998,244(5):71—84. doi: 10.1016/S0375-9601(98)00312-0
[4]	Judd Kevin,Alistair Mess. Embedding as a modeling problem[J].Phys D,1998,120(4):273—286. doi: 10.1016/S0167-2789(98)00089-X
[5]	Schroer Christian G,Sauer Tim,Ott Edward,et al.Predicting chaotic most of the time from embeddings with self-intersections[J].Phys Rev Lett,1998,80(7):1410—1412. doi: 10.1103/PhysRevLett.80.1410
[6]	Chon Ki H.Detection of chaotic determinism in time series from randomly forced maps[J].Phys D,1997,99(5):471—486. doi: 10.1016/S0167-2789(96)00159-5
[7]	Kitoh Satoshi,Kimura Mahito,Mori Takao,et al. A fundamental bias in calculating dimension from finite data sets[J].Phys D,2000,141(10):171—182. doi: 10.1016/S0167-2789(00)00050-6
[8]	马军海,陈予恕,刘曾荣.动力系统实测数据的非线性混沌模型重构[J].应用数学和力学，1999,20(11):1128—1134.
[9]	马军海，陈予恕.低维混沌时序非线性动力系统的预测方法及其应用研究[J].应用数学和力学，2001,22(5):441—448.
[10]	马军海，陈予恕，刘曾荣.不同随机分布的相位随机化对实测数据影响的分析研究[J].应用数学和力学，1998,19(11):955—964.
[11]	马军海，陈予恕，刘曾荣.动力系统实测数据的Lyapunov指数的矩阵算法[J].应用数学和力学，1999,20(9):919—927.
[12]	马军海，陈予恕.混沌时序相空间重构的分析和应用研究[J].应用数学和力学，2000,21(11):1237—1245.
[13]	马军海，陈予恕，刘曾荣. 动力系统实测数据的非线性混沌特性的判定[J]. 应用数学和力学1998,19(6):481—488.