共查询到19条相似文献,搜索用时 125 毫秒
1.
为解决过程神经元网络不能直接输入离散样本的问题,提出基于样条插值函数的离散过程神经网络训练算法。首先,将离散过程样本按采样点分段,在采样区间内分别构造样本和权值的分段样条函数;然后,计算样本函数和权函数的乘积在采样区间上的积分,并将此积分值提交给网络的隐层过程神经元;最后,在输出层计算网络输出。分别采用一次、二次、三次样条函数,设计了三种不同的算法。实验结果表明:一次样条计算效率高,逼近能力差;三次样条计算效率低,但逼近能力好;二次样条在计算效率和逼近能力两方面都比较理想。因此,二次样条函数是离散过程神经网络的较好选择。 相似文献
2.
基于三次样条函数拟合的过程神经元网络训练 总被引:3,自引:2,他引:3
过程神经元网络的提出为大样本识别问题开辟了新途径,但其训练方法目前主要基于权函数正交基展开。这种方法基函数个数选取目前尚无理论依据。提出了基于三次样条函数拟合的过程神经元网络训练方法。首先将样本函数、过程神经元权函数的离散化数据拟合成分段表示的三次样条函数,然后计算样本样条函数与权值样条函数乘积在给定采样区间上的积分,并将此积分值提交给网络的过程隐层神经元,输出层由普通神经元组成。三次样条函数具有很好的光滑性、可积性、阶数低、参数少等优点,有效地简化了网络的时空聚合运算。实验表明该方法是可行的。 相似文献
3.
提出了一种基于最优分段函数逼近的过程神经网络学习算法。将网络时变输入信号和连接权在一定精度下表示为分段函数的拟合形式,根据最小均方误差准则,构建 PN N基于函数基展开的训练算法。选择低阶分段函数作为基函数,利用其良好的柔韧逼近和光滑可导性质,快速实现网络待定参数对函数样本的自适应学习。网络训练中,只需迭代调整分段函数的连接系数,可有效减少模型中的参数冗余、提高PNN对实际问题的建模能力。 相似文献
4.
5.
为了有效解决传统人工神经网络对于时变函数的聚类问题,以及提高在大样本下网络的学习和泛化能力,提出了基于离散余弦变换的传统人工神经网络动态过程聚类方法。通过离散余弦变换将样本函数降维映射到由对应余弦参数所张成的模式特征空间,满足了传统人工神经网络对输入样本的要求,使传统人工神经网络实现动态过程的聚类成为可能。给出了实现算法,分析了计算复杂度,并使用基本竞争型人工神经网络对特征样本向量进行聚类,实验结果表明该方法是正确、有效的。与过程人工神经网络相比,该方法具有运算简单、物理意义明确等优点。 相似文献
6.
基于函数正交基展开的过程神经网络学习算法 总被引:27,自引:1,他引:27
过程神经网络的输入和连接权均可为时变函数,过程神经元增加了一个对于时间的聚合算子,使网络同时具有时空二维信息处理能力.该文在考虑过程神经网络对时间聚合运算的复杂性的基础上,提出了一种基于函数正交基展开的学习算法.在网络输入函数空间中选择一组适当的函数正交基,将输入函数和网络权函数都表示为该组正交基的展开形式,利用基函数的正交性.简化过程神经元对时间的聚合运算.应用表明,算法简化了过程神经网络的计算复杂度,提高了网络学习效率和对实际问题求解的适应性.以旋转机械故障诊断问题和油藏开发过程采收率的模拟为例验证了算法的有效性. 相似文献
7.
为提高神经网络的逼近能力,提出一种基于序列输入的神经网络模型及算法。模型隐层为序列神经元,输出层为普通神经元。输入为多维离散序列,输出为普通实值向量。先将各维离散输入序列值按序逐点加权映射,再将这些映射结果加权聚合之后映射为隐层序列神经元的输出,最后计算网络输出。采用Levenberg-Marquardt算法设计了该模型学习算法。仿真结果表明,当输入节点和序列长度比较接近时,模型的逼近能力明显优于普通神经网络。 相似文献
8.
张代远 《模式识别与人工智能》2004,17(4)
讨论了输入层和输出层为恒等变换神经元,隐层为非线性变换神经元的三层神经网络的统计灵敏度,给出了权值和输入样本的加性扰动灵敏度计算的理论公式.当输入扰动较小时,通过选择多种不同的隐层神经元函数,按照本文理论结果计算出来的灵敏度数值与仿真实验的结果十分吻合,说明了本文理论分析结果的正确性.当输入扰动较小时,隐层取不同函数时灵敏度的值是不同的,但数值基本上稳定不变.对于某些隐层函数,当输入扰动值较大时,网络的灵敏度数值迅速下降,但是网络输出的平均相对误差却迅速增加.在求灵敏度时,网络的权值是用代价函数为0的精确算法求得的. 相似文献
9.
自组织过程神经网络及其应用研究 总被引:5,自引:1,他引:5
针对与时间过程有关的模式分类问题,提出了一种自组织过程神经元网络模型.网络由输入层和竞争层组成,其输入和连接权可为与时间有关的函数,输入层结点与竞争层结点实行全互连接.网络提取输入函数所隐含的过程式模式特征,并对其进行自组织,在竞争层将分类结果表现出来.为简化计算,在输入空间中引入函数正交基,将输入函数和网络权函数表示为正交基的展开形式,利用基函数的正交性,使网络权函数的调整非时变化.给出了竞争学习和有教师示教两种学习算法,并以石油地质中沉积微相识别问题为例证明了模型和算法的有效性. 相似文献
10.
概率神经网络(PNN)-径向基网络的重要变形,它的学习速发快,很适合于故障检测问题,但是当网络输入样本过大时,网络的计算就会很复杂,计算速度就会很缓慢.本文提出用主元分析(PCA)对过程数据进行降维,然后将处理过的数据作为网络输入,这样使网络的计算速度得到了提高.最后将提出的方法用于田纳西伊斯曼过程(Tennessee... 相似文献
11.
Louis Wang Gang Feng 《IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics》2004,34(1):682-686
This paper presents a new H infinity controller design method for the discrete time fuzzy systems based on the piecewise Lyapunov functions. The basic idea of the proposed approach is to construct the controller for the fuzzy systems in such a way that a discrete time piecewise Lyapunov function can be used to establish the global stability with H infinity-disturbance attenuation performance of the resulting close loop fuzzy control systems. It is shown that the control laws can be obtained by solving a set of linear matrix inequalities (LMIs) that is numerically tractable with commercially available software. Numerical example is given to demonstrate the advantage of the proposed method. 相似文献
12.
Gang FENG 《控制理论与应用(英文版)》2003,1(1):28-34
This paper presents an H∞ controller design method for pieccwise discrete time linear systems based on a piecewise quadratic Lyapunov function. It is shown that the resulting closed loop system is globally stable with guaranteed H∞ perfomiance and the controller can be obtained by solving a set of bilinear lnatrLx inequalities. It has been shown that piecewise quadratic Lyapunov functions are less conservative than the global qnadnmc Lyapunov functions. A simulation example is also given to illustrate the advantage of the proposed approach. 相似文献
13.
Dieter D. Pfaffinger 《Computer Methods in Applied Mechanics and Engineering》1980,24(3):269-286
Dynamic excitations in the form of stationary random processes with normal distribution are completely defined by their power spectral and cross spectral density functions. The stationary response of a linear structure to such excitations will also consist of random processes with normal distribution. In a modal formulation the statistical quantities of all output processes are obtained from modal covariance matrices. The elements of these matrices represent integrals which are usually evaluated numerically. In lightly damped structures, however, the integrand shows pronounced peaks. Thus small integration steps may be necessary for accurate results. In the applications the spectral density functions are conveniently described by discrete values and piecewise polynomial interpolation. The elements of the modal covariance matrices can then be evaluated analytically. For lightly damped structures this method is much more effective than numerical integration and maintains full accuracy in the modal properties of the structural model. The accuracy and efficiency of the method is illustrated by a numerical example. 相似文献
14.
The operational matrix of piecewise linear polynomial functions is derived. Then the problems of analysis and parameter identification are studied on this piecewise linear polynomial basis. The algorithms proposed are analogous to those derived for the orthogonal functions. The piecewise polynomial basis has certain advantages and convenience over other orthogonal functions. Illustrative examples are given for demonstration. The satisfactory results are obtained owing to the rapid convergence property of the piecewise linear polynomial functions. 相似文献
15.
《Fuzzy Systems, IEEE Transactions on》2005,13(4):531-543
This paper presents an observer based$H_infty$ output feedback synthesis method for discrete time fuzzy dynamic systems based on a piecewise Lyapunov function. The basic idea of the approach is to design an observer based piecewise linear output feedback control law to guarantee the global stability with$H_infty$ performance of the resulting closed-loop fuzzy control systems. It is shown that the controller parameters can be obtained by solving a set of linear matrix inequalities (LMIs) that are numerically feasible with commercially available software. Application to control chaotic systems is given to illustrate the effectiveness and advantages of the proposed method. 相似文献
16.
BMDP program for piecewise linear regression 总被引:1,自引:0,他引:1
T Nakamura 《Computer methods and programs in biomedicine》1986,23(1):53-55
Piecewise linear regression has potentially broad applications in medical data analysis as well as other types of regression. Various kinds of algorithms have been proposed for finding optimum piecewise linear regressions. This paper presents a BMDP program for obtaining near optimum piecewise linear regression equations. An idea intrinsic to the method is that restricting parameter space to a discrete set makes the difficult problems become standard problems. Any software having the variable selection feature in the multiple linear regression can be used to apply the method. 相似文献
17.
This article presents a switched model reference adaptive controller for discrete‐time piecewise linear systems. In the spirit of the work by Landau in the late seventies, proof of asymptotic stability of the closed‐loop error system is obtained, recasting its dynamics as a feedback system and showing the feedforward and the feedback paths are both passive. The challenge is that both paths can be piecewise linear. Numerical results show excellent performance of the proposed controller even in the face of sudden variations of the plant parameters. Copyright © 2012 John Wiley & Sons, Ltd. 相似文献
18.
Anne Gelb 《Journal of scientific computing》2007,30(3):409-440
Spectral series expansions of piecewise smooth functions are known to yield poor results, with spurious oscillations forming
near the jump discontinuities and reduced convergence throughout the interval of approximation. The spectral reprojection
method, most notably the Gegenbauer reconstruction method, can restore exponential convergence to piecewise smooth function
approximations from their (pseudo-)spectral coefficients. Difficulties may arise due to numerical robustness and ill-conditioning
of the reprojection basis polynomials, however. This paper considers non-classical orthogonal polynomials as reprojection
bases for a general order (finite or spectral) reconstruction of piecewise smooth functions. Furthermore, when the given data
are discrete grid point values, the reprojection polynomials are constructed to be orthogonal in the discrete sense, rather
than by the usual continuous inner product. No calculation of optimal quadrature points is therefore needed. This adaptation
suggests a method to approximate piecewise smooth functions from discrete non-uniform data, and results in a one-dimensional
approximation that is accurate and numerically robust.
相似文献
19.
针对一类不确定离散时滞分段系统,研究了广义H2稳定性分析和带有时滞的状态反馈弹性控制器设计问题.通过构造适当的离散分段二次李亚普诺夫函数,利用分段二次李亚普诺夫稳定性理论,给出了对于所有的容许参数不确定性,闭环系统是广义H2稳定的充分条件;在此基础上,基于线性矩阵不等式(LMI)处理方法,提出了带有时滞的状态反馈弹性控制器增益阵的设计方案.仿真实例验证了所提方法的有效性. 相似文献