Development of modular electrical systems

Modular systems provide the ability to achieve product variety through the combination and standardization of components. A methodology that combines system modeling, integration analysis, and optimization techniques for development of modular systems is presented. The approach optimizes integration and interactions of system elements and creates functional and physical modules for the electrical system. The Hatley/Pirbhai methodology (1987) is used for modeling functional requirements of a system. The model defines system interfaces (interactions) to support its functions. Once the interactions among functions are identified, an incidence matrix of the interfaces is developed. A clustering algorithm is developed to identify clusters in the incidence matrix, group the functions, and create modules. A Hatley/Pirbhai architecture model is developed to represent modular system design. A detailed discussion on the importance of system modeling in design of modular systems and on the constraints that limit development of modular vehicle systems is also presented. The approach presented is systematic and can be used to support product development and decision-making in engineering design     

边界积分法及连接算法分析任意腔体的散射

本文利用边界积分法分析二维任意腔体的散射,给出一种基于微波网络原理的连接算法,将腔体分为几段,分别用积分方程法计算每段的广义导纳矩阵,然后利用连接算法将各段连接,得到整个腔体的口径导纳矩阵,最后由广义网络原理求解腔体的等效磁流及后向散射场。本文方法可作为一种机辅设计算法。     

广义网络法分析复杂孔缝耦合问题

本文利用广义网络法结合连接算法分析复杂孔缝耦合问题.首先根据孔缝的结构及填充特点将其内腔体分为适当的几段,利用边界元法分别计算每段的广义导纳矩阵,再借助连接算法将各段连接起来得到整个孔缝的口径导纳矩阵,最后由广义网络法求解孔缝的口径磁流、散射及传输场.该方法不仅在计算效率方面取得了较大突破,也使复杂填充孔缝的分析得到很大简化.     

导电平面上三维任意腔体的散射分析

本文利用边界积分法及连接算法分析导电平面上的三维腔体散射。在引入广义导纳矩阵后,可将腔体分为几段,分别用积分方程法计算每段的广义导纳矩阵。然后利用连接算法得到整个腔体的口径导纳矩阵。最后由广义网络原理求解腔体的口径等效磁流及后向散射场。本文方法极大地缓解了计算机内存对腔体尺寸的限制,提高了分析效率,可作为一种机辅设计算法。     

Electromagnetic scattering from three-dimensional cavities via aconnection scheme

The electromagnetic scattering from arbitrary three-dimensional cavities is presented. To alleviate computational constraints for three-dimensional problems, a connection scheme is developed based on microwave network theory. This scheme allows the cavity to be divided into sections and each section to be analyzed independently of the rest of the cavity. Each section of the cavity is represented by a generalized admittance matrix which if formulated via a boundary-integral equation approach. Using the concept of input and load admittance, the aperture admittance matrix of the cavity can be derived by cascading the admittance matrices of individual sections. Once the cavity aperture admittance matrix is obtained, the aperture electric field and the backscattered field are found by the standard generalized network formulation. Numerical results are compared against modal solutions of regularly shaped cavities with good agreement. This connection scheme leads to a reduction in computational resources, especially for cavities with one dimension much larger than the other two     

A quasi-Newton adaptive algorithm for generalized symmetriceigenvalue problem

We first recast the generalized symmetric eigenvalue problem, where the underlying matrix pencil consists of symmetric positive definite matrices, into an unconstrained minimization problem by constructing an appropriate cost function. We then extend it to the case of multiple eigen-vectors using an inflation technique. Based on this asymptotic formulation, we derive a quasi-Newton-based adaptive algorithm for estimating the required generalized eigen-vectors in the data case. The resulting algorithm is modular and parallel, and it is globally convergent with probability one. We also analyze the effect of inexact inflation on the convergence of this algorithm and that of inexact knowledge of one of the matrices (in the pencil) on the resulting eigenstructure. Simulation results demonstrate that the performance of this algorithm is almost identical to that of the rank-one updating algorithm of Karasalo (1986). Further, the performance of the proposed algorithm has been found to remain stable even over 1 million updates without suffering from any error accumulation problems     

阶梯波导的传输矩阵研究

用模式匹配方法分析H面阶梯波导的TE模式,首先将各区域的场写为x函数矩阵与z函数矩阵以及本区域的场系数向量相乘的形式,然后对相邻区域的x函数矩阵进行加权积分得到矩阵形式的匹配方程组。不仅在表达形式上简洁紧凑,而且能够方便的推导出考虑高次凋落模的传输矩阵。通过传输矩阵求解场,未知数个数成倍减小,运算量显著降低。数值例子验证了方法的正确性。     

距离决策下的模糊聚类集成模型

模糊聚类是近年来使用的一类性能较为优越的聚类算法,但该类算法对初始聚类中心敏感且对边界样本的聚类结果不够准确。为了提高聚类准确性、稳定性,该文通过联合多个模糊聚类结果,提出一种距离决策下的模糊聚类集成模型。首先,利用模糊C均值(FCM)算法对数据样本进行多次聚类,得到相应的隶属度矩阵。然后,提出一种新的距离决策方法,充分利用得到的隶属度关系构建一个累积距离矩阵。最后,将距离矩阵引入密度峰值(DP)算法中,利用改进的DP算法进行聚类集成以获取最终聚类结果。在UCI机器学习库中选择9个数据集进行测试,实验结果表明,相比经典的聚类集成模型,该文提出的聚类集成模型效果更佳。     

A new approach for analysis of resonant structures based on thespatial finite-difference and temporal differential formulation

This paper presents a new procedure for analyzing resonant structures using the spatial finite-difference and temporal differential formulation. Unlike the conventional finite-difference time-domain methods, the finite-differences are only enforced in the spatial domain for Maxwell's equations. The time-domain differentials of Maxwell's equations are kept, resulting in a system of first-order differential equations. In consequence, a resonant structure problem can be formulated in the eigenvalue problem form and resonant modes are obtained by solving the corresponding eigenvalue problem directly. It is shown that the coefficients of the matrix for the eigenvalue problem can be simply obtained from the finite-difference time-domain formulation. As a result, an efficient alternative way of using the finite-difference time-domain approach to solve the resonant structure problems is presented. The algorithm is applied to metallic waveguide structures and the numerical results agree well with those from other techniques     

基于相似性传播的天波雷达多路径量测聚类

电离层多层结构特性使得天波雷达(OTHR)与目标之间存在多条信号传播路径,进而可能对单目标产生多路径量测。该文考虑了天波雷达多路径量测聚类问题,其需要同时对多路径量测进行电离层传播路径辨识和聚类。由于天波雷达量测模型假设1个目标通过1种电离层传播路径至多产生1个量测,因此需要考虑多路径聚类约束。该文将相似性传播聚类扩展到多路径约束模型,并提出一种新的多路径相似性传播聚类算法。该算法通过构建多路径量测聚类的概率图模型,将聚类问题转化为概率图模型隐变量的推断问题,采用最大和置信传播算法近似求解聚类变量的最大后验概率。算法优点包括可以自动识别聚类团数目,单次消息传播的时间复杂度为量测个数和传播路径个数乘积的平方。仿真实验分析表明,所提算法较多路径多假设聚类算法具有更好的聚类性能。     

A new eigenvector selection strategy applied to develop spectral clustering

Spectral methods are strong tools that can be used for extraction of the data’s structure based on eigenvectors of constructed affinity matrices. In this paper, we aim to propose some new measurement functions to evaluate the ability of each eigenvector of affinity matrix in data clustering. In the proposed strategy, each eigenvector’s elements are clustered by traditional fuzzy c-means algorithm and then informative eigenvectors selection is performed by optimization of an objective function which defined based on three criterions. These criterions are the compactness of clusters, distance between clusters and stability of clustering to evaluate each eigenvector based on considering the structure of clusters which placed on. Finally, Lagrange multipliers method is used to minimize the proposed objective function and extract the most informative eigenvectors. To indicate the merits of our algorithm, we consider UCI Machine Learning Repository databases, COIL20, YALE-B and PicasaWeb as benchmark data sets. Our simulation’s results confirm the superior performance of the proposed strategy in developing spectral clustering compared to conventional clustering methods and recent eigenvector selection based algorithms.     

数据挖掘中一种改进的谱组合聚类算法

组合聚类(EC)是解决数据挖掘问题的关键手段之一,但现有的EC方法较少考虑可能破坏聚类结构的各种噪声,降低了聚类性能。为此,提出一种改进的谱组合聚类(ISEC)方法。将聚类问题建模为输入的多个基本分区(BPs)派生的共协矩阵的图分割问题;ISEC方法学习得到共协矩阵的低秩表示,并在共协矩阵上进行谱聚类,提高聚类性能;最后采用增强拉格朗日乘数法进行优化求解,获得最终的聚类结果。在多个真实数据集上的仿真实验结果表明,ISEC方法的聚类性能优于目前的大多数聚类方法。     

Electromagnetic transmission through inhomogeneously filled slotsin a thick conducting plane-arbitrary incidence

The penetration of an arbitrarily incident electromagnetic wave through a slot filled with inhomogeneous material in a thick conducting plane is analyzed. The solution is obtained via a combined finite-element method/method of moments algorithm based on the generalized network formulation. The discretization of the generalized network formulation is performed via the method of moments. The finite-element method is then used to compute the fields within the inhomogeneous interior cavity region, leading to the construction of the interior aperture admittance matrix. It is shown that with the use of entire domain basis functions, the construction of the aperture admittance matrices is computationally efficient. Furthermore, this method is attractive since it preserves the sparsity of the finite-element method matrix, reducing computational memory requirements. Some examples of the penetration of inhomogeneously filled slots of various cross sections are presented     

基于特征加权和非负矩阵分解的多视角聚类算法

为了在多视角聚类过程中同时考虑特征权重和数据高维性问题,提出一种基于特征加权和非负矩阵分解的多视角聚类算法(Multiview Clustering Algorithm based on Feature Weighting and Non-negative Matrix Factorization,FWNMF-MC).FWNMF-MC算法根据每个视角中每个特征在聚类过程中的重要性,自动赋予不同的权值.通过将每个视角空间中的特征矩阵分解为基矩阵与系数矩阵的乘积,将多视角数据从高维空间映射到低维空间.为了有效利用每个视角信息挖掘聚簇结构,最大化每个视角在低维空间的一致性.最后实验结果表明FWNMF-MC算法的聚类效果明显优于已有的4种有代表性的多视角聚类算法.     

Underdetermined BSS Based on <Emphasis Type="Italic">K</Emphasis>-means and AP Clustering

Underdetermined blind source separation (UBSS) is a hard problem to solve since its mixing system is not invertible. The well-known “two-step approach” has been widely used to solve the UBSS problem and the most pivotal step is to estimate the underdetermined mixing matrix. To improve the estimation performance, this paper proposes a new clustering method. Firstly, the observed signals in the time domain are transformed into sparse signals in the frequency domain; furthermore, the linearity clustering of sparse signals is translated into compact clustering by normalizing the observed data. And then, the underdetermined mixing matrix is estimated by clustering methods. The K-means algorithm is one of the classical methods to estimate the mixing matrix but it can only be applied to know the number of clusters in advance. This is not in accord with the actual situation of UBSS. In addition, the K-means is very sensitive to the initialization of clusters and it selects the initial cluster centers randomly. To overcome the fatal flaws, this paper employs affinity propagation (AP) clustering to get the exact number of exemplars and the initial clusters. Based on those results, the K-means with AP clustering as initialization is used to precisely estimate the underdetermined mixing matrix. Finally, the source signals are separated by linear programming. The experimental results show that the proposed method can effectively estimate the mixing matrix and is more suitable for the actual situation of UBSS.     

RPCL算法的理论发展和应用

在传统的聚类分析中,通常需要针对给定的数据选择出正确或合理的类别数,否则算法无法得到理想的聚类分析结果。当采用竞争学习（Competitive Learning, CL）算法进行聚类分析时也面临着同样的问题。然而,一般数据集中实际聚类个数（或竞争单元个数）的推断与选择却是一个十分困难的问题。为了解决这一难题,对手惩罚竞争学习（Rival Penalized Competitive Learning, RPCL）算法建立了一种有效的思想和方法。它通过预设较大的聚类个数,在竞争学习中引入了对手惩罚的机制,自动地选择出正确的聚类中心与个数,并将多余的聚类中心排除到无穷点或远离数据的地方。这种独特的思想和方法为聚类分析开辟了一条崭新的途径。本文将深入分析RPCL算法的理论发展,包括产生的根源及其思想、理论基础、在不同情况下的推广和变式,并且总结了RPCL算法在各个领域中的应用。     

Formulation and solution of structured total least norm problemsfor parameter estimation

The total least squares (TLS) method is a generalization of the least squares (LS) method for solving overdetermined sets of linear equations Ax≈b. The TLS method minimizes ∥[E|-r]∥_F, where r=b-(A+E)x, so that (b-r)∈Range (A+E), given A∈C^m×n, with m⩾n and b∈C^m×1. The most common TLS algorithm is based on the singular value decomposition (SVD) of [A/b]. However, the SVD-based methods may not be appropriate when the matrix A has a special structure since they do not preserve the structure. Previously, a new problem formulation known as structured total least norm (STLN), and the algorithm for computing the STLN solution, have been developed. The STLN method preserves the special structure of A or [A/b] and can minimize the error in the discrete L_p norm, where p=1, 2 or ∞. In this paper, the STLN problem formulation is generalized for computing the solution of STLN problems with multiple right-hand sides AX≈B. It is shown that these problems can be converted to ordinary STLN problems with one right-hand side. In addition, the method is shown to converge to the optimal solution in certain model reduction problems. Furthermore, the application of the STLN method to various parameter estimation problems is studied in which the computed correction matrix applied to A or [A/B] keeps the same Toeplitz structure as the data matrix A of [A/B], respectively. In particular, the L₂ norm STLN method is compared with the LS and TLS methods in deconvolution, transfer function modeling, and linear prediction problems     

基于自适应增量学习的时间序列模糊聚类算法

针对现存可用于时间序列的增量式模糊聚类算法往往需要设置多个控制参数的问题,本文提出了一种基于自适应增量学习的时间序列模糊聚类算法.该算法首先继承上一次聚类得到的簇结构信息以初始化当前聚类进程,然后在无需设置参数的情况下自适应地搜索当前数据块中的离群样本,并自动从离群样本创建新簇,最后检查空簇识别标识确定是否需要移除部分簇以保证后续聚类过程的效率.实验结果表明所提算法对等长和不等长时间序列均具有良好的聚类准确性及运行效率.     

一种有效的启发式聚类算法

本文讨论了一种利用确定性退火技术的启发式聚类算法。它把聚类问题看作一物理系统。通过求解一系列随温度变化的自由能函数的全局极小来得到聚类问题的最优解。算例表明，对传统聚类算法无能为力的几种聚类问题，该算法都得到了比较满意的结果。     

一种模糊加权的改进层次聚类算法研究

传统层次聚类算法中经常会遇到合并点和分裂点选择的问题,一旦一组对象被合并或者分裂,下一步的处理将在新生成类上进行,已做处理不能撤销,这样有可能导致低质量的聚类结果.针对这个问题,文中提出了一种模糊加权层次聚类改进算法,每次分层聚类时先计算对象属于这个类可靠度,然后和阀值进行比较,当可靠度小于阀值时重新确定对象的归属类,这样就解决了上述问题.最后通过实验验证,该算法确实可行有效.