可约割集和可约回路及其在网络分析中的应用

本文以梯形网络为讨论对象,提出可约分的独立割集和可约分的独立回路的概念。从而导出结构分析法,解决了常规网络分析法中某些研究得尚不透彻的问题。结构分析的结果与状态空间分析法所得结果完全一致,而且使用简便。最后给出了实例。     

On the reliability of a number of s-p reducible graphs

Let G=(V,E) be a graph whose edges may fail with known probabilities and let K a subset of V be specified. The overall reliability of G, denoted by R(G), is the probability that all vertices in K=V communicate with each other. We have two types of graphs, s-p reducible and s-p complex, depending on whether after series-arallel reductions the result is a single edge or not. A number of s-p reducible graphs are presented and expressions that evaluate their overall reliability are introduced.     

结合ICA和复数神经网络的二麦阵列盲源分离技术

为了降低语音信号盲源分离算法的延时,提高其准确性和稳定性,本文结合传统盲源分离技术和深度神经网络的优势,提出了一种基于ICA独立分量分析和复数神经网络的二麦阵列盲源分离技术。本文将复数递归神经网络和独立分量分析方法有机融合,提出一种基于时频域的双通道复数神经网络,同时解决了独立分量分析中的排列问题。所提方法利输入混合信号利用复数域神经网络计算初始化分离矩阵,神经网络输出采用复数域形式,利用复数学习标签估计复数矩阵,然后采用独立分量分析方法获得目标分离矩阵。实验数据表明,所提方法相较于其它独立分量分析方法提高了盲源分离的实时性和准确性。      

Automatic generation of starting values for the simulation ofmicrowave oscillators by frequency domain techniques

A new approach to the start-up problem inherent to the large-signal analysis of autonomous circuits in the frequency domain is presented. By insertion of a simple network, depending on one parameter, the oscillator is damped to the stability limit where a linear analysis yields good results. The steady state of the undamped oscillator is then obtained by a continuation method corresponding to the successive removal of the damping network. With this procedure the degenerate solution may be excluded in a straightforward manner     

单模真空场中两个耦合二能级原子纠缠的时间演化特性

研究了单模真空场中两个耦合二能级原子纠缠特性.运用密度矩阵方法,得到两能级原子约化密度矩阵.借助于共生纠缠度,研究两个耦合二能级原子纠缠的时间演化规律,讨论原子间偶极作用对原子间纠缠的影响.结果表明:原子间的偶极作用导致纠缠度的减少和时间演化周期的增加.     

基于BP神经网络-Monte Carlo法的结构可靠性分析

提出通过人工神经网络拟合极限状态函数的方法来解决结构可靠性问题。根据多层神经网络映射存在定理,对于任何在闭区间内的一个连续函数都可以用含有一个隐含层的BP网络来逼近。应用此定理,通过人工神经网络拟合极限状态方程,借助神经网络的函数映射关系产生大量的极限状态函数值,作为下一步的分析数据。此过程并不像Monte Carlo法时每一点都做确定性计算,因而达到减少计算工作量的目的。该方法仅采用Monte Carlo法随机抽样的思路,对大范围的数据进行概率分析,通过概率分析得到极限状态函数值的均值和标准差,以便求得结构系统的可靠性指标,进行结构系统可靠性分析。     

三能级原子同时存在辐射和吸收时的量子相干性

量子位和环境的相互作用,将引起量子位能量耗散或相对位相改变,最终导致量子位由相干叠加态退化为混合态或纯态,这种现象被称为量子消相干.针对三能级原子同时存在辐射和吸收时的情况,基于量子力学和Kraus算子理论,利用约化密度矩阵方法,分析了三能级原子密度矩阵元的演化规律,从而研究该系统的量子相干特性.结果表明:初态处于相干叠加态的三能级原子,由于辐射和吸收,可能演化成相干叠加态、未完全混合态、完全混合态、纯态.     

Congested Banyan network analysis using congested-queue states andneighboring-queue effects

A Banyan network analysis technique is presented which more accurately models a congested network than other reported techniques. The analysis is based on the observation that a full queue (within the switching modules making up the network) causes traffic to back up. For a short time after becoming full, the free space in the queue is limited to no more than one slot. A queue in such a condition is called congested. Because of blocking the arrival rate to a congested queue is higher; this tends to maintain congestion. The arrival rate to a congested queue's dual is lower, as is the service rate for queues feeding the congested queue. These effects are captured in the analysis by using a state model that codes the condition of a queue and its neighbors. From a queue's state it can be determined if the queue, its dual, or a next-stage queue to which it connects is congested, as well as the number of items in the queue. Network throughput computed with the model is closer to that obtained from simulations than other Banyan analyses appearing in literature, including those designed to model congestion. Further, the queue-occupancy distributions are much closer than other analyses, suggesting that the analysis better models conditions in congested Banyan networks     

Reliability evaluation of a communication system considering a multistate model

The methods most often suggested for reliability evaluation of communication systems/flow networks consider two states, namely the good state and the bad state. This paper considers multistate modelling of the network elements. The states which permit a flow less than the maximum assigned capacity are assumed as parallel and mutually exclusive elements in the network. A method is proposed for reliability evaluation of a communication network considering multistate modelling of the network elements. The proposed method has been implemented on a computer and results obtained using the proposed method are compared with the existing methods for the two state model conditions.     

Reliability computation of multistage interconnection networks

A general algorithm that generates a reliability expression for these measures is presented. The reliability is analyzed under the assumption of statistical independence of failure. Nonetheless, failure dependencies can also be solved. This method is appropriate for analyzing other computer networks. To model networks with switching elements (SEs), the SE is considered as good or bad. The complex fault models are described. The analytic method is implemented by a Pascal program. The advantages of the proposed method are: (1) flexibility-this technique can handle statistically independent and dependent failures within the network, while failure dependencies use the Lam-Li model; this technique is easily extendable for the analysis of M-input N-output multistage interconnection networks; (2) generality-this technique derives the reliability of networks from the points of view of terminal reliability, broadcasting, and network reliability; and (3) cost effectiveness-the necessary system information can be obtained at very little additional cost, once the connection equation of the network has been obtained     

A novel approach for generating boundary condition independent compact dynamic thermal networks of packages

A novel approach is proposed for generating compact dynamic thermal models of packages, independent on boundary conditions. This approach is based on the novel definition of Robin's boundary condition independent dynamic thermal network and on the application of a novel multivariate moment matching method. The resulting compact thermal models are more accurate than compact models obtained by means of previous methods.     

一种基于RefineDet网络和霍夫变换的高速铁路接触网吊弦状态多尺度检测方法

针对高速铁路接触网吊弦的状态检测问题,该文提出一种基于RefineDet网络和霍夫变换的吊弦多尺度定位与识别方法。通过设计RefineDet网络的粗调和精调模块对吊弦整体结构进行定位,采用霍夫变换锁定吊弦中部吊悬线所在直线,并利用旋转因子沿直线方向提取吊悬线区域;以吊悬线区域代替吊弦结构整体区域送入分类网络进行训练,通过所建立的多尺度吊弦状态检测模型,实现吊弦状态的精确识别。实验结果表明,吊弦定位模型的准确率达95.3%以上;霍夫变换可排除无效区域对吊弦状态识别的干扰,提高分类网络的训练速度,吊弦状态识别模型准确率达97.5%以上。     

Simulation of AC spacecraft power systems

A technique is presented for modeling and simulating AC spacecraft power systems by combining all component models into an overall system model. Each component in the spacecraft power system is treated as a two-port network. A state model is written for each two-port network with the port voltages as the inputs. Each component state model is solved independently using the state transition matrix approach and assuming that its inputs are constant. The inputs to all components are then calculated using network analysis principles. As an example, a 20 kHz system is simulated using this approach, and the results are compared with those of a SPICE2 simulation     

An impedance measurement method for double-gap klystron cavity

A new method has been developed for measuring the impedance of a two-gap cavity used in high-power klystrons. The principle is based on network analysis. The cavity under test is considered as a microwave network. The two gaps of the cavity and its output terminal are referred to as three ports. We then can use an impedance matrix to characterize this system, and the six independent impedance parameters can be found by measuring the input impedance seen from the output waveguide when the gaps are in different conditions; viz., either open, shorted, or perturbed. The gap impedance then can be deduced therefrom. It is shown that there are three impedances for a double-gap cavity instead of one for a single-gap cavity. Another problem dealt with here is how to evaluate the capacitance introduced by the perturbation. A few typical experimental results are presented.     

Circular grid solution for Laplace's equation in two dimensions

Often in the analysis of electromagnetic field problems reducible to Laplace's equation in two dimensions, solution is required for only a few points. A numerical method which is applicable to the solution of such problems, provided that the boundary is circular, is given. Excellent results have been obtained also for some cases with piecewise-linear boundaries.     

一种基于神经网络的卡尔曼滤波改进方法

卡尔曼(Kalman)滤波是一种基于最小方差估计的递推式滤波方法,它要求信号的状态模型是已知的,这就限制了它在实际中的应用。利用神经网络的良好的非线性映射能力对实际系统进行系统辨识,可以获得符合精度要求的系统状态方程,很大程度改进了卡尔曼滤波的效果。相对于一些经典的卡尔曼滤波改进算法,这种方法具有应用范围广和数学建模简单易行的优点。将神经网络与卡尔曼滤波相结合的方法用于图像复原实验,结果表明,该方法具有可行性和有效性。     

Statistical analysis technique on Ad Hoc network topology dynamic characteristics: Markov stochastic process

On the basis of analysis on the scene files of mobility models in Ad Hoc network, the paper presents a network topology snapshots capturing method to obtain the Ad Hoc network topology architecture at any moment. Through analyzing on the Ad Hoc network topology snapshots, some dynamic characteristic parameters of Ad Hoc network, such as the number of network topology in steady state or unsteady state appearing during a certain time, as well as the durative time of network topology in steady state or unsteady state, could be obtained statistically. Furthermore, the probability of the network topology invariability and variability event could be predicated by adopting the discrete time and continuous time Markov stochastic process theory. The simulation result shows that the statistical analysis technique on Ad Hoc network topology dynamic characteristic not only is effective, but also has the general attribute, which could be used in the statistical analysis technique on Ad Hoc network topology dynamic characteristic under any mobility models.     

Hadoop下基于贝叶斯网络的气象数据挖掘研究

基于朴素贝叶斯的分类器是气象数据挖掘中比较传统的方法,但由于算法要求各属性相互独立,预测精度无法达到要求,且在处理海量数据时算法计算效率受到制约,对此提出一种Hadoop平台下基于离散贝叶斯网络的数据挖掘改进算法。算法不要求属性之间相互独立,且充分结合Hadoop平台适应处理大数据的优点,利用海量数据分析地面气象因素与温度之间的相关性,并由此选取预测因子来训练贝叶斯网络分类器模型,以达到预测温度的目的。实验结果表明,算法不但预测精度明显高于目前短期气候预测中采用的朴素贝叶斯算法,而且极大地提高了运算效率。     

Algorithm for analysis of general linear active networks

The steady state of the most general active linear network can be obtained by methods using very frequently high-order matrices which contain a great number of zeros. A method is proposed which avoids this, the order of the `active? matrix being equal to the number of controlled sources introduced in the network.     

基于匹配博弈的车辆异构网络选择算法

摘要：针对车辆在异构网络覆盖环境下的网络接入选择问题,基于双边匹配博弈中稳定匹配的相关概念,提出一种车辆异构网络选择博弈模型,通过用户侧与网络侧的双向选择得到异构网络选择的稳定匹配。算法首先利用层次分析法获得相应权重分配以给出满意度函数,然后将网络侧的一对多匹配转化为一对一匹配问题,在稳定匹配条件下构建多目标优化模型,并求解最优匹配结果。相关性能分析及仿真实验表明,基于稳定匹配的博弈模型相对于传统决策模型,能够使网络和用户侧双方的满意度最优化,从而得到双向网络选择的均衡状态。