一种基于关联维数的混沌时间序列分形插值建模.pdfVIP

th Proceedings of the 25 Chinese Control Conference 7-11 August, 2006, Harbin, Heilongjiang 一种基于关联维数的混沌时间序列分形插值建模 付冲，曹英禹，张振川 1. 东北大学信息科学与工程学院, 沈阳 110004 E-mail: fuchong@ise.neu.edu.cn 摘 要: 提出了一种基于关联维数的复杂系统混沌时间序列的分形插值建模算法，解决了传统随机分布模型由于 增加不必要的自由度从而不能准确描述系统时空演化规律的缺陷.以Internet流量时间序列为例，在相空间重构 的基础上，利用自相关函数确定了系统的最佳时延，根据饱和嵌入维数求得系统的关联维数.最后利用关联维数 确定系统的Hurst指数，通过分数布朗运动插值算法对混沌时间序列进行了精确的重构，从而建立一种可精确重 演混沌系统时空变化规律的模型. 关键词: 混沌系统, 关联维数, Hurst指数, 分形插值 A Relational Dimension Based Fractal Interpolation Algorithm for Chaotic Time Serials Modeling Chong Fu, Ying-yu Cao, Zhen-chuan Zhang School of Information Science and Engineering, Northeastern University, Shenyang 110004, China E-mail: fuchong@ise.neu.edu.cn Abstract: This paper proposes a relational dimension based fractal interpolation algorithm for complex system chaotic time serials modeling, which resolves the disadvantage of the traditional random distribution model that the system space-time evolution rules can not be precisely described due to the unnecessary added degree of freedom. The Internet traffic is used as analysis object, the phase space is reconstructed and the system optimal delay is determined by auto correlation function. The system relational dimension is calculated out by saturation embedded dimension analysis and based on which the Hurst exponent of the system is gained. The chaotic time serials are reconstructed by using fractal Brown motion interpolation algorithm, which establishes a nonlinear model for precisely describing a chaotic system evolution process. Key Words: chaotic system, relational dimension, Hurst