优化Hammerstein模型的电磁脉冲效应建模仿真

针对采用Hammerstein模型描述电磁脉冲效应时模型参数确定困难的问题,提出一种基于优化Hammerstein模型的非线性电磁脉冲效应建模仿真方法。该方法利用高阶累积量求解模型线性延迟阶数,采用赤池信息准则（Akaike Information Criterion,AIC）优化模型非线性阶次,两者共同确定出一个参数最少、精度较高的最优模型,并以瞬态抑制器的电磁脉冲效应建模为例验证了该方法。     

自适应类别的层次高斯混合模型遥感影像分割

为了实现自动确定类别数的高精度遥感影像分割,提出一种自适应类别的层次高斯混合模型（Hierarchical Gaussian Mixture Model,HGMM）遥感影像分割算法.提出算法采用多个高斯混合模型加权和定义HGMM,用于建模具有非对称,重尾和多峰等复杂特性的影像统计模型.采用期望最大化算法（Expectation Maximization,EM）求解模型参数.为了实现自动确定类别数,采用贝叶斯信息准则（Bayesian Information Criterion,BIC）求解最优类别数,其中惩罚项采用加权像素数定义.为了验证提出算法可行性和有效性,对模拟和全色遥感影像进行分割实验,并对分割结果进行定性和定量分析.结果表明HGMM具有准确建模复杂统计分布的能力,提出算法具有高精度和高效率,同时可自动确定最优类别数.     

移动通信网络评价准则与解决方案

面向未来巨大差异化的业务、多样的个性化体验要求,论述了关键性能指标（KPI）、关键质量指标（KQI）、业务质量（QoS）、业务体验（QoE）及关键经营指标（KBI）的内在关系以及从KPI、KQI和KBI出发各自可能的最优设计方法及潜在的挑战和技术研究方向,这是一个层层递进的统一整体。最后指出,目前的公平性指标是一种后验的“经验”指导,没有产生具体的反馈闭环机制以形成对设计准则的明确指导。而从KBI角度的牵引,可能是可选的终结性、归一化的评价准则。需要指出的是,要从深度和广度上做好这些工作,需要跨学科、跨领域的深入研究,需要跨网络不同节点间的密切协同、跨层设计。     

A characterization of the zero-free region of the Riemann zeta function and its applications

We characterize the zero-free regions of a class of functions (including the Riemann zeta function) in half-planes in terms of closures of ranges of the corresponding multiplication operators on Hardy spaces. We give an explicit characterization of these closures. As applications, we obtain a weaker version of the Nyman–Beurling–Báez-Duarte criterion, and provide some investigations on a problem relating to the Riemann hypothesis proposed by Báez-Duarte et al. [Adv. Math. 149 (2000) 130-144].     

大规模数据下子抽样模型平均估计理论

随着信息时代的来临,如何从海量数据中快速、有效地挖掘有用信息是目前面临的新挑战.子抽样方法作为大规模数据分析的有效工具,已经受到国内外学者的广泛关注.不过,传统的子抽样方法通常没有考虑到模型的不确定性.当模型假设不正确时,后面的统计推断将会出现偏差,甚至导致错误的结论.为了解决该问题,文章利用频率模型平均的方法构建了子抽样模型平均估计(简称SSMA估计).理论上,文章证明了SSMA估计是全部数据下模型平均估计的一个渐近无偏且相合的估计.另外,我们基于Hansen (2007)的Mallows模型平均方法提出了SSMA估计的权重选择准则,并证明了方差已知和未知时权重估计的渐近最优性.在这些理论性质的研究中,文章同时考虑了模型和抽样设计带来的双重随机性.最后,数值分析进一步说明了所提出方法的有效性.     

Some contributions to selection and estimation in the normal linear model

We study the choice of the quantity in the FPE criterion for selecting a member of a class of normal linear models having an orthogonal structure. Two approaches are discussed, namely fixing the maximal estimation risk at a prescribed level and using minimax regret. Estimation of the risk actually achieved and an illustrative example are also discussed.This research was supported by the FRD of South Africa.     

Analysis of mechanical behavior in cold-drawing deformation of polymers

When stretched under uniaxial stress, ductile polymers usually exhibit unstable plastic deformation, which embodies two phases: (a) yielding with the formation of a neck and (b) cold-drawing with the propagation of necking shoulders. The mechanical state associated with this deformation behavior is analyzed. The discussion is divided into three parts. The first part is a general treatment of the constitutive function of flow stress in the plastic state, in which a series of relations among various characterizing parameters are formulated. The second part provides three mechanical criteria for necking deformation and propagation of necking shoulders: the condition of unstable plastic deformation requiring $ D_P = - \left( {\partial {{\ln \dot \varepsilon } \mathord{\left/ {\vphantom {{\ln \dot \varepsilon } {\partial \varepsilon }}} \right. \kern-\nulldelimiterspace} {\partial \varepsilon }}} \right)_P < 0 $ the stabilizing deformation mode, which requires $ \gamma _p = \left( {{{dD_p } \mathord{\left/ {\vphantom {{dD_p } {d\varepsilon }}} \right. \kern-\nulldelimiterspace} {d\varepsilon }}} \right)_P > 0 $ and the obvious localization of unstable plastic deformation. The third part describes a mathematical model which can be used in calculations to fit the contour of the necking shoulder. This model is developed according to rational considerations for the relation of In $ \dot \varepsilon $ to ε. Experimental data on PE rod specimens are well fitted by this model. © 1993 John Wiley & Sons, Inc.     

Time-domain optimization of input signals for distributed-parameter systems identification

The problem of the optimum signal choice for the purpose of distributed-parameter systems identification is considered. In contrast to earlier contributions to this topic, the time-domain approach leads to a nonconvex optimization problem. This difficulty is overcome by introducing an auxiliary convex problem, which is proved to have the same solution as the initial one. This approach allows one to obtain necessary and sufficient optimality conditions for the initial problem, which are then used to find analytical solutions for a wide class of systems. The so-calledD-Optimality criterion is used here, together with constraints on the signal energy.The author expresses his sincere thanks to the anonymous referees for their helpful comments and suggestions.     

基于知识库的系统安全评估方法

在系统非正常状态和正常状态这两种运行状况下,使用nmon等系统监测工具采集了代表系统运行状态的指标数据,对获取的指标数据进行了向量化操作,构造了可以反映系统运行状态的多维行为特征.采用类内、类间评价距离准则对指标数据进行了有效特征提取,构建了能够反映系统运行状态的行为特征知识库.然后利用行为特征知识库的数据构造了决策树,以判断系统所处状态是否安全.     

Stable adaptive channel estimation method under impulsive noise environments

Channel estimation is one of the key technologies for ensuring reliable wireless communications under impulsive noise environments. This paper studies robust adaptive channel estimation methods for mitigating harmful impulsive noises, which are described as alpha‐stable (α ‐stable) distribution models. Traditional adaptive channel estimation using the second‐order statistics based least mean square (SOS‐LMS) algorithm does not perform well under α ‐stable noise environments, even though it was considered one of attractive approaches for estimating channels in the case of Gaussian noises. Unlike the traditional SOS‐LMS algorithm, in this research, we propose a stable sign‐function‐based LMS algorithm, which can mitigate the impulsive noises. Specifically, we first construct the cost function with minimum ℓ ₁‐norm error criterion and then derive the updating equation of the proposed algorithm. Compared with the traditional SOS‐LMS, the effectiveness of the proposed algorithm is validated via Monte Carlo simulations in various α ‐stable noise scenarios. Copyright © 2015 John Wiley & Sons, Ltd.