基于MSC双声路SAW器件的设计仿真平台

利用模块化设计思想,在Matlab编程环境下实现了基于多条耦合器双声路声表面波器件的设计仿真平台。为了提高MSC双声路SAW器件的性能,该平台将多条耦合器、分裂指及假指设计方案融合到器件设计中。多条耦合器不仅实现了两个变迹换能器传递函数的乘积,而且消除了体声波对器件性能的影响。假指消除了通过变迹换能器引起的波前畸变。分裂指可以最小化换能器内部的声波反射问题。通过一个频率为203.559MHz的基于多条耦合器双声路声表面波器件的设计和实际测试结果的对比分析,证明了该设计仿真平台不仅可以进行器件版图结构的设计和模拟,而且设计平台所得到的版图和数据可以用于器件的掩膜版制作和器件封装尺寸参考。另外,该仿真设计平台可以很好地实现对器件的脉冲特性、频率响应特性的模拟和仿真。     

大规模软件系统的无标度特性与演化模型

在软件工程中，常用类图来描述类之间的关系. 以软件系统网为研究对象，通过对Sun和IBM公司提供的大规模软件系统进行实证分析，发现Java软件系统网的度分布是无标度分布，标度指数γ≈2.5. 在软件系统网的演化过程中，除加点之外，还存在边的添加、边的随机移除与边的重连等局部事件. 由此建立了软件系统演化模型. 由该模型演化生成的网络，其度分布服从幂律分布. 实际应用与数值仿真验证了该模型的有效性. 关键词： 软件系统 复杂网络 度分布 无标度     

Poor rich demarcation of Matthew effect on scale-free systems and its application

In a scale-free network,only a minority of nodes are connected very often,while the majority of nodes are connected rarely. However,what is the ratio of minority nodes to majority nodes resulting from the Matthew effect In this paper,based on a simple preferential random model,the poor-rich demarcation points are found to vary in a limited range,and form a poor-rich demarcation interval that approximates to k/m ∈ [3,4]. As a result,the (cumulative) degree distribution of a scale-free network can be divided into three intervals: the poor interval,the demarcation interval and the rich interval. The inequality of the degree distribution in each interval is measured. Finally,the Matthew effect is applied to the ABC analysis of project management.     

基于超声相控阵的卫星适配器缺陷检测方法研究

根据卫星适配器的缺陷检测要求,设计并实现了一种基于超声相控阵的微小缺陷的超声相控阵自动检测方法.该方法以超声相控阵的检测理论为核心,选择合适的检测方案,搭建超声相控阵自动检测系统.以供借鉴.     

The complexity nature of large-scale software systems

In software engineering, class diagrams are often used to describe the system's class structures in Unified Modelling Language (UML). A class diagram, as a graph, is a collection of static declarative model elements, such as classes, interfaces, and the relationships of their connections with each other. In this paper, class graphs are examined within several Java software systems provided by Sun and IBM, and some new features are found. For a large-scale Java software system, its in-degree distribution tends to an exponential distribution, while its out-degree and degree distributions reveal the power-law behaviour. And then a directed preferential-random model is established to describe the corresponding degree distribution features and evolve large-scale Java software systems.