基于径向基神经网络预测的混沌时间序列嵌入维数估计方法

根据Takens定理,研究了混沌时间序列相空间重构嵌入维数的选取问题.提出了基于径向基函数神经网络预测模型性能的嵌入维数估计方法,即根据嵌入维数与混沌时间序列预测模型性能的变化关系来确定嵌入维数.通过对几种典型混沌动力学系统的数值验证,结果表明该方法能够确定出合适的相空间重构嵌入维数. 关键词： 混沌 相空间重构 嵌入维数 预测     

混沌时序相空间重构参数确定的信息论方法

根据信息论基本原理，研究了混沌时间序列相空间重构参数延迟时间和嵌入维数的选取.提出了用符号分析的方法计算互信息函数，确定出延迟时间，在此基础上，提出了一种估计嵌入维数的信息论方法，即根据重构向量条件熵随向量维数的变化关系来确定嵌入维数，通过对几种典型混沌动力学系统的数值验证，结果表明该方法能够确定出合适的相空间重构嵌入维数. 关键词： 混沌 相空间重构 互信息 条件熵 符号分析     

混沌时间序列重构相空间参数选取研究

基于重构相空间的延迟时间和嵌入维数这两个参数的选取不相关的观点,提出用互信息函数法确定延迟时间后,用CAO方法来确定嵌入维数的新思路.通过对几种典型的混沌动力学系统的数值验证,结果表明该方案能够确定出相空间重构的有效延迟时间和最佳嵌入维数.该方法能够从时间序列中有效地重构原系统的相空间,是混沌信号识别的一种有效途径.     

螺旋桨鸣音的混沌动力特性研究

利用混沌动力学方法研究螺旋桨鸣音信号时间序列,估计时间序列的相空间重构最佳参数,并提出其具有混沌动力特性,分析了系统拓扑维数的边界和生成系统所必须独立变量的个数,还计算分析了重构相空间中吸引子轨迹随时间演化的发散情况。分析计算结果表明:螺旋桨鸣音信号时间序列可以选取最佳延迟时间t_D=1、最小嵌入维数d_E=8进行相空间重构,其混沌吸引子的关联维数为5.1579、最大Lyapunov指数为0.0771,此研究结果可以为螺旋桨鸣音现象的进一步研究提供理论基础。      

基于条件熵扩维的多变量混沌时间序列相空间重构

提出一种多变量混沌时间序列相空间重构的条件熵扩维方法.首先使用互信息法求解每个变量的时间延迟,其次按条件熵最大原则逐步扩展相空间的嵌入维数,使得重构坐标从低维到高维的转换保持较强的独立性,最终的重构相空间具有较低的冗余度,为多变量时间序列的预测构造了有效的模型输入向量.通过对几个经典多变量混沌时间序列进行数值实验,结果表明该方法比单变量预测和已有多变量预测方法具有更好的预测效果,说明了该重构方法的有效性. 关键词： 多变量混沌时间序列 相空间重构 条件熵 神经网络预测     

混沌时间序列的支持向量机预测

根据混沌动力系统的相空间延迟坐标重构理论,基于支持向量机的强大的非线性映射能力, 建立了混沌时间序列的支持向量机预测模型,并在统计学习理论的基础上采用最小二乘方法来训练预测模型,利用该模型对嵌入维数与模型的均方根误差的关系进行了探讨.最后利用Mackey-Glass时间序列和变参数的Ikeda 时间序列对该模型进行了验证,结果表明,该预测模型能精确地预测混沌时间序列,而且在混沌时间序列的嵌入维数未知时也能取得比较好的预测效果.这一结论预示着支持向量机是一种研究混沌时间序列的有效方法. 关键词： 混沌时间序列 支持向量机 最小二乘法     

基于径向基函数神经网络的混沌时间序列相空间重构双参数联合估计

鉴于径向基函数(RBF)神经网络模型在非线性预测方面的优良性能, 提出了利用该预测模型对混沌时间序列相空间重构的两个关键参数——延迟时间和嵌入维数进行联合估计的方法, 并以客观的评价指标为依据给出其最优估计值. 以Lorenz系统为例进行数值分析, 得到RBF单步及多步预测模型中嵌入维数和延迟时间的最佳参数估计值, 并在原模型中对估计值进行校验. 结果表明, 该方法可以有效地估计出嵌入维数和延迟时间, 从而显著提高预测精度.     

横掠管束周期性充分发展对流换热的混沌分析

本文利用混沌理论分析了横掠管束周期性充分发展对流换热的非稳定性问题,即通过速度U的时间序列的重构相空间计算出关联维数D2,并通过时间序列分析了该非线性动力系统的功率谱特性。分析结果表明,本文所研究的横掠管束周期性充分发展对流换热系统在所给出的控制参数Re=937.7下出现的非稳定性问题属于混沌现象。系统的整体状态可用奇怪吸引子来描述,当延迟时间选择为5,该时间序列的重构相空间的嵌入维数增至5时,该吸引子的分维数趋于定值1.63。     

短期风速时间序列混沌特性分析及预测

针对短期风速时间序列的预测问题进行了研究. 首先通过0-1混沌测试法确定短期风速时间序列具有混沌特性. 采用相空间重构技术, 利用C-C算法确定延迟时间, G-P 算法确定嵌入维数. 然后提出一种参数在线修正的最小二乘支持向量机预测模型, 采用改进的粒子群算法进行预测模型中参数的优化. 最后通过仿真对比实验表明提出的预测方法在预测精度、预测误差、预测效果方面都要优于其他常见的预测方法, 证明该预测方法是有效的.     

基于最优相空间重构参数的空调系统负荷预测

空调系统负荷非线性变化,单独利用人工神经网络方法进行预测,受到样本数量少等的限制,精度一般都比较低。空调系统负荷时间序列包含了参与空调系统动态变化的全部变量的信息,可以利用相空间重构技术提取和恢复出系统原来的规律。重构后的相空间具有与实际动力系统相同的几何性质,这样就可以大大增加样本数。相空间重构受嵌入维数m和延迟时间τ的影响,确定这两个重构参数的最佳数值是非常重要的。建立了基于相空间重构技术的神经网络空调负荷预测模型,在嵌入维数m和延迟时间τ分别取不同的值时,利用这个模型对相同时间段的空调负荷进行预测,选择预测误差最小时对应的参数为最佳相空间重构参数,对应的预测值为最佳预测结果。结果表明,这个方法有很好的效果,具有简化计算复杂性等特点。     

Research on the chaos recognition method based on differential entropy

Phase space reconstruction is the first step to recognizing the chaos from observed time series. On the basis of differential entropy, this paper introduces an efficient method to estimate the embedding dimension and the time delay simultaneously. The differential entropy is used to characterize the disorder degree of the reconstructed attractor. The minimum value of the differential entropy corresponds to the optimum set of the reconstructed parameters. Simulated experiments show that the original phase space can be effectively reconstructed from time series, and the accuracy of the invariants in phase space reconstruction is greatly improved. It provides a new method for the identification of chaotic signals from time series.     

Macroscopic Time Evolution and MaxEnt Inference for Closed Systems with Hamiltonian Dynamics

MaxEnt inference algorithm and information theory are relevant for the time evolution of macroscopic systems considered as problem of incomplete information. Two different MaxEnt approaches are introduced in this work, both applied to prediction of time evolution for closed Hamiltonian systems. The first one is based on Liouville equation for the conditional probability distribution, introduced as a strict microscopic constraint on time evolution in phase space. The conditional probability distribution is defined for the set of microstates associated with the set of phase space paths determined by solutions of Hamilton’s equations. The MaxEnt inference algorithm with Shannon’s concept of the conditional information entropy is then applied to prediction, consistently with this strict microscopic constraint on time evolution in phase space. The second approach is based on the same concepts, with a difference that Liouville equation for the conditional probability distribution is introduced as a macroscopic constraint given by a phase space average. We consider the incomplete nature of our information about microscopic dynamics in a rational way that is consistent with Jaynes’ formulation of predictive statistical mechanics, and the concept of macroscopic reproducibility for time dependent processes. Maximization of the conditional information entropy subject to this macroscopic constraint leads to a loss of correlation between the initial phase space paths and final microstates. Information entropy is the theoretic upper bound on the conditional information entropy, with the upper bound attained only in case of the complete loss of correlation. In this alternative approach to prediction of macroscopic time evolution, maximization of the conditional information entropy is equivalent to the loss of statistical correlation, and leads to corresponding loss of information. In accordance with the original idea of Jaynes, irreversibility appears as a consequence of gradual loss of information about possible microstates of the system.     

Information Theory Applied to Parameter Inference of Pake Doublet Spectra

Magnetic resonance absorption lineshapes can have subtle dependencies on the model parameters that specify the lineshape. To quantify how the model parameters influence the lineshape, it is useful to study simple model systems for which analytical expressions are available. We propose that information theory is a useful tool to quantify how well model parameters may be inferred from a noisy signal. Information theory also allows us to assess the importance of missing parameters from an incomplete model. We do this by monitoring the magnitude of a partition function determined from a suitably defined probability mass function as the model parameters are varied. The optimum parameter set makes the partition function a maximum, which establishes a computable criterion for determining the best model parameter set. Given the availability of a partition function, one may define thermodynamic functions such as the entropy. The optimum parameter set in this interpretation corresponds to the state of maximum entropy. In this work, we observe that at sufficiently low signal to noise ratio, the entropy landscape has no clear maximum, while a related quantity, the Fisher information, always has a clear minimum at the optimum parameter set. The qualitative information we are able to gather from the entropy landscapes is also difficult to assess when the parameters are far from their optimum values, at least for the model system studied here.     

非平衡统计信息理论

阐述了以表述信息演化规律的信息（熵）演化方程为核心的非平 衡统计信息理论.推导出了 Shannon信息（熵）的非线性演化方程，引入了统计物理信息并 推导出了它的非线性演化方程.这两种信息（熵）演化方程一致表明：统计信息（熵）密度 随时间的变化率是由其在坐标空间（和态变量空间）的漂移、扩散和减损（产生）三者引起 的.由此方程出发，给出了统计信息减损率和统计熵产生率的简明公式、漂移信息流和扩散 信息流的表达式，证明了非平衡系统内的统计信息减损（或增加）率等于它的统计熵产生（ 或减少）率、信息扩散与信息减损同时 关键词： 统计信息（熵）演化方程 统计信息减损率 统计熵产 生率 信息（熵）流 信息（熵）扩散 动态互信息     

Dynamic statistical information theory

In recent years we extended Shannon static statistical information theory to dynamic processes and established a Shannon dynamic statistical information theory, whose core is the evolution law of dynamic entropy and dynamic information. We also proposed a corresponding Boltzmman dynamic statistical information theory. Based on the fact that the state variable evolution equation of respective dynamic systems, i.e. Fokker-Planck equation and Liouville diffusion equation can be regarded as their information symbol evolution equation, we derived the nonlinear evolution equations of Shannon dynamic entropy density and dynamic information density and the nonlinear evolution equations of Boltzmann dynamic entropy density and dynamic information density, that describe respectively the evolution law of dynamic entropy and dynamic information. The evolution equations of these two kinds of dynamic entropies and dynamic informations show in unison that the time rate of change of dynamic entropy densities is caused by their drift, diffusion and production in state variable space inside the systems and coordinate space in the transmission processes; and that the time rate of change of dynamic information densities originates from their drift, diffusion and dissipation in state variable space inside the systems and coordinate space in the transmission processes. Entropy and information have been combined with the state and its law of motion of the systems. Furthermore we presented the formulas of two kinds of entropy production rates and information dissipation rates, the expressions of two kinds of drift information flows and diffusion information flows. We proved that two kinds of information dissipation rates (or the decrease rates of the total information) were equal to their corresponding entropy production rates (or the increase rates of the total entropy) in the same dynamic system. We obtained the formulas of two kinds of dynamic mutual informations and dynamic channel capacities reflecting the dynamic dissipation characteristics in the transmission processes, which change into their maximum—the present static mutual information and static channel capacity under the limit case where the proportion of channel length to information transmission rate approaches to zero. All these unified and rigorous theoretical formulas and results are derived from the evolution equations of dynamic information and dynamic entropy without adding any extra assumption. In this review, we give an overview on the above main ideas, methods and results, and discuss the similarity and difference between two kinds of dynamic statistical information theories.     

Entanglement and localization of a two-mode Bose-Einstein condensate

A simple second quantization model is used to describe a two-mode Bose-Einstein condensate (BEC), which can be written in terms of the generators of a SU(2) algebra with three parameters. We study the behavior of the entanglement entropy and localization of the system in the parameter space of the model. The phase transitions in the parameter space are determined by means of the coherent state formalism and the catastrophe theory, which besides let us get the best variational state that reproduces the ground state energy. This semiclassical method let us organize the energy spectrum in regions where there are crossings and anticrossings. The ground state of the two-mode BEC, depending on the values of the interaction strengths, is dominated by a single Dicke state, a spin collective coherent state, or a superposition of two spin collective coherent states. The entanglement entropy is determined for two recently proposed partitions of the two-mode BEC that are called separation by boxes and separation by modes of the atoms. The entanglement entropy in the boxes partition is strongly correlated to the properties of localization in phase space of the model, which is given by the evaluation of the second moment of the Husimi function. To compare the fitness of the trial wavefunction its overlap with the exact quantum solution is evaluated. The entanglement entropy for both partitions, the overlap and localization properties of the system get singular values along the separatrix of the two-mode BEC, which indicates the phase transitions which remain in the thermodynamical limit, in the parameter space.     

Quantum confinement in an asymmetric double‐well potential through energy analysis and information entropic measure

Localization of a particle in the wells of an asymmetric double‐well (DW) potential is investigated here. Information entropy‐based uncertainty measures, such as Shannon entropy, Fisher information, Onicescu energy, etc., and phase‐space area, are utilized to explain the contrasting effect of localization‐delocalization and role of asymmetric term in such two‐well potentials. In asymmetric situation, two wells behaves like two different potentials. A general rule has been proposed for arrangement of quasi‐degenerate pairs, in terms of asymmetry parameter. Further, it enables to describe the distribution of particle in either of the deeper or shallow wells in various energy states. One finds that, all states eventually get localized to the deeper well, provided the asymmetry parameter attains certain threshold value. This generalization produces symmetric DW as a natural consequence of asymmetric DW. Eigenfunctions, eigenvalues are obtained by means of a simple, accurate variation‐induced exact diagonalization method. In brief, information measures and phase‐space analysis can provide valuable insight toward the understanding of such potentials.     

Random-injection-based two-channel chaos with enhanced bandwidth and suppressed time-delay signature by mutually coupled lasers: Proposal and numerical analysis

Simultaneous bandwidth(BW) enhancement and time-delay signature(TDS) suppression of chaotic lasing over a wide range of parameters by mutually coupled semiconductor lasers(MCSLs) with random optical injection are proposed and numerically investigated. The influences of system parameters on TDS suppression(characterized by autocorrelation function(ACF) and permutation entropy(PE) around characteristic time) and chaos BW are investigated. The results show that, with the increasing bias current, the ranges of parameters(detuning and injection strength) for the larger BW(> 20 GHz) are broadened considerably, while the parameter range for optimized TDS(< 0.1) is not shrunk obviously.Under optimized parameters, the system can simultaneously achieve two chaos outputs with enhanced BW(> 20 GHz)and perfect TDS suppression. In addition, the system can generate two-channel high-speed truly physical random number sequences at 200 Gbits/s for each channel.     

Analysis of systems based on entropy and information characteristics

An approach is proposed for studying various systems based on the concept of entropy potentials of their parameters. In this approach, any state of a system can be described quantitatively using the values of entropy potentials of corresponding parameters. With such an approach, the states and the evolution processes in the systems being analyzed can be mapped onto the coordinate space of the “information field” of the system, which makes it possible to simplify the solution of a number of applied problems.