Adaline对权扰动敏感性的近似计算

主要讨论Adaline对权扰动敏感性的计算.鉴于Adaline的输入和输出的不连续性,敏感性定义为Adaline对于所有可能输入在权值发生扰动的情况下输出发生变化的概率.借助超球面模型和解析几何的技术,给出一个近似计算Adaline敏感性的方法.在输入维数足够大的情况下,该方法优于以往的其它方法.它在牺牲少量计算精度的情况下,极大地降低了计算复杂度,使得敏感性更具实用价值.     

基于离散随机过程建立锂电池模型的研究

本文首先通过锂电池实验深入研究了电池的两个基本特性,恢复效应和速率容量效应.然后依据离散随机电池模型理论,模拟出跟踪电池基本行为的离散过程,通过仿真和锂电池放电实验进行了模型验证,结果显示模拟过程与理论推导相吻合,从而验证了模型的正确性.     

离散时间系统中随机谐振研究

摘要：基于一种新的信噪比的定义讨论了在四种经典噪声下在离散时间系统中的随机谐振现象。当输入信号是阈上信号时,噪声总是在恶化信号的传输,即随着噪声强度增加,输出信噪比呈现单调递减的趋势。然而当输入信号是阈下信号时。适当的噪声能够引起系统最优反应,即随着噪声强度的增加,输出信噪比达到一个最大峰值,也即存在明显的随机谐振现象。随着非线性系统中阈值的增加,随机谐振现象的明显度降低,当最佳噪声强度在增加时,输出信噪比的最优值却在减少。文中结果也说明,基于新的信噪比的定义是可以用来度量一个离散时间系统的随机谐振现象,且随机谐振对噪声具有一定的鲁棒性,拓广了随机谐振在信息传输领域的应用。     

基于动态观测的随机离散事件系统故障诊断

近年来,离散事件系统故障诊断研究引起国内外学者广泛关注.鉴于此,研究动态观测下随机离散事件系统的故障诊断.首先引入一种动态观测,使事件的可观测性随着系统的运行而动态变化;然后分别对基于动态观测的随机离散事件系统的单故障可诊断性和模式故障可诊断性进行形式化;最后通过构造相应的诊断器,分别得到关于单故障可诊断性和模式故障可...     

离散随机系统的包含原理

提出线性时变离散随机系统的包含原理.根据系统包含的定义,给出系统状态观测与反馈控制闭环系统的约束条件和聚集条件,使具有重叠结构和模型降阶系统的反馈设计或LQG设计有充分的选择.     

基于离散小波变换和随机森林的轴承故障诊断研究

针对不同工况下数据特征选择困难和单一分类器在滚动轴承故障诊断中识别率较低等问题,提出了一种基于离散小波变换和随机森林相结合的滚动轴承故障诊断方法。该方法首先利用离散小波变换分解振动信号,得到n层近似系数;然后创新性地采用sigmoid熵构造出n维特征向量,sigmoid熵能较好地提取非平稳信号的特征,提高诊断准确率;最后采用随机森林对滚动轴承不同故障信号进行分类。实验采用西储凯斯大学轴承数据中心网站提供的轴承数据,与传统分类器(KNN和SVM)以及单个分类回归树CART进行对比分析,结果表明该方法具有更好的诊断效果。     

广义离散随机线性系统的马尔可夫估计

本文讨论了带有脉冲模的广义离散随机线性系统Exk=Φxk-1+гwk-1,yk=Hxk+vk的马尔可夫状态估计问题,得到了这个估计的递推算法.     

随机中断情境下的离散型设施选址问题研究

针对传统的供应链设施选址模型大多是基于静态背景下的确定性选址问题研究,而较少考虑中断风险因素的情况,基于随机中断情境,建立了混合整数规划模型表述的设施可靠性选址问题模型,采用拉格朗日松弛算法进行求解。通过构建的算例,求解了问题模型的最优解并验证了该算法的求解性能。     

基于随机扰动技术的多级矢量量化器优化设计

提出了一种高效的随机扰动多级矢量量化编码方案，实验证明该算法不仅能降低计算复杂度和码书存储量，而且在较高压缩比下可获得较好的图象恢复。     

改进的非线性离散系统自适应数据驱动控制

基于同步扰动随机近似的算法, 控制器选取为一个函数逼近器, 并在这里被确定为神经网络. 控制算法中使用了自适应的参数估计, 明显改善了控制性能, 同时也给出了相应的收敛性分析. 最后, 新型的控制算法被应用到了解决非线性离散系统的跟踪控制问题中, 并通过仿真比较结果, 充分验证了这种自适应数据驱动控制策略的可行性和有效性.     

Approximate synthesis of optimal control over quasilinear stochastic differential equations with a small parameter and Poisson perturbations

The authors obtain successive approximations to the optimal control over a quasilinear system of stochastic functional-differential equations with Poisson disturbances and a quadratic quality functional. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 3, pp. 39–45, May–June 2008.     

Parameter identification technique for multivariate stochastic systems

This paper considers the problem of obtaining accurate estimates of multivariate systems with reasonable computations. To avoid the structural identification problem which is associated with multivariate systems, we observe the system by a linear combination of the outputs. The two stage least square method is employed to estimate the model parameters. An optimum combination of the outputs is obtained such that the parameter estimates have the least asymptotic generalized variance. Computer simulations are provided to illustrate the usefulness of the proposed method.     

Method of stochastic sensitivity synthesis in a stabilisation problem for nonlinear discrete systems with incomplete information

A problem of the synthesis of the assigned probabilistic distribution near the equilibrium of stochastic nonlinear discrete-time system with incomplete state information is considered. We construct a static feedback regulator which provides a required stochastic sensitivity of this equilibrium. It is shown that this problem is reduced to the solution of some quadratic matrix equations. The solvability of these quadratic equations is analysed, and attainability sets are described. In the two-dimensional case, two variants of the control input are discussed and compared. These general results are applied to the suppression of large-amplitude oscillations around the equilibria of the stochastically forced Henon model with noisy observations. We show how suggested control technique, by minimising the stochastic sensitivity, allows us to suppress chaos and provide a structural stabilisation.     

Efficient simulation of hierarchical stochastic activity network models

Stochastic extensions to Petri nets have gained widespread acceptance as a method for describing the dynamic behavior of discrete-event systems. Both simulation and analytic methods have been proposed to solve such models. This paper describes a set of efficient procedures for simulating models that are represented as stochastic activity networks (SANs, a variant of stochastic Petri nets) and composed SAN-based reward models (SBRMs). Composed SBRMs are a hierarchical representation for SANs, in which individual SAN models can be replicated and joined together with other models, in an iterative fashion. The procedures exploit the hierarchical structure and symmetries introduced by the replicate operation in a composed SBRM to reduce the cost of future event list management. The procedures have been implemented as part of a larger performance-dependability modeling package known asUltraSAN, and have been applied to real, large-scale applications. This work was supported in part by the Digital Equipment Corporation Faculty Program: Incentives for Excellence.     

基于灵敏度分析法的ELM剪枝算法

针对极端学习机(ELM)网络结构设计问题,提出基于灵敏度分析法的ELM剪枝算法.利用隐含层节点输出和相对应的输出层权值向量,定义学习残差对于隐含层节点的灵敏度和网络规模适应度,根据灵敏度大小判断隐含层节点的重要性,利用网络规模适应度确定隐含层节点个数,删除重要性较低的节点.仿真结果表明,所提出的算法能够较为准确地确定与学习样本相匹配的网络规模,解决了ELM网络结构设计问题.     

Full-state perturbation analysis of discrete event dynamic systems

Discrete event dynamic systems are studied within the framework of perturbation analysis in this paper. Perturbation is extended from the event times only to both event times and queue lengths. An approximate technique, full-state perturbation analysis (PA), is developed as an extension of the PA approach. Full-state PA is able to deal with problems involving queue length perturbations which often defy existing PA methods, while it still retains all the advantages of existing PA. Full-state PA is used to calculate the throughput sensitivity to the number of customers in closed queueing networks and the throughput sensitivity to routing change. Numerical examples are given. Experimental results verify the validity and accuracy.This work is supported in part by the National High Technology Project and by Southeast University Research Funds for Young Teachers.     

Sensitivity analysis and optimization of stochastic Petri nets

In this paper we define the notion of controlled stochastic Petri net (CSPN), which is a stochastic Petri net with controlled parameters and performance indices. Specifically, transition times and/or conflict resolution rules can depend on controlled parameters and transition times can have arbitrary distribution functions. A method for computing statistical estimates of performance indices and their gradients (sensitivities) with respect to controlled parameters is described. This method, which needs only one simulation of a CSPN, is considerably superior to conventional finite differences both in terms of precision and required amount of simulation and is based on likelihood ratio/score function approach, other possibilities based on extensions of infinitesimal perturbation analysis are outlined. These gradient estimates are used in stochastic optimization algorithms to obtain the optimal value of the aggregated performance function of the CSPN. A combined optimization and simulation tool is developed which includes approaches to the gradient estimation mentioned above. The numerical experiments presented in this paper confirm the efficiency of the proposed techniques.