主动网络中基于多代理系统的流量均衡*

主动网络流量均衡控制ATEC是一种在主动网络环境中,基于多代理MA系统的网络性能控制机制。该方法将多路径路由和资源分配相结合,利用非线性规划启发式算法——多路径等性能策略,实现网络流量和带宽资源的均衡配置,从而避免或减少网络拥塞的发生,达到提高网络性能的目的。着重介绍了ATEC的多代理MA系统实现,包括基本算法、一般模型、代理设计和合作模型的设计,并给出了仿真结果。     

主动网络中CPU和带宽资源分配的公平性

传统的包调度方案能实现单一资源分配的公平性,但不能直接应用于主动网络中的主动节点上,因为主动节点包含多种相互关联的资源。提出一种能实现在主动节点中公平分配资源的包调度算法。仿真实验结果显示,该算法能在主动流之间实现CPU和带宽资源的有效公平分配。     

SAFE: An Efficient Feature Extraction Technique

This paper proposes an efficient window-based semi-automatic feature extraction technique which uses simulated annealing for minimizing the energy of an active contour within a specified image region. The energy is computed based on a chamfer image, in which pixel values are a function of distance to image edges. A user places a number of control points close to the feature of interest. B-spline fitted to these points provides an initial approximation of the contour. A window containing both the initial contour and the feature of interest is considered. The contour with minimum energy inside the window provides the final delineation. Comparison of the performance of the proposed algorithm with traditional snake, a popular feature extraction technique based on energy minimization, demonstrates the superiority of the SAFE technique. Received 18 August 1999 / Revised 25 October 2000 / Accepted in revised form 8 December 2000     

Active sitting with backrest support: Is it feasible?

Abstract

Ergonomics science recommends office chairs that promote active sitting to reduce sitting related complaints. Since current office chairs do not fulfill this recommendation, a new chair was developed by inverting an existing dynamic chair principle. This study compares active sitting on the inverted chair during a simulated computer-based office task to two existing dynamic office chairs (n?=?8). Upper body stability was analysed using Friedman ANOVA (p?=?.01). In addition, participants completed a questionnaire to rate their comfort and activity after half a working day.

The inverted chair allowed the participants to perform a substantial range of lateral spine flexion (11.5°) with the most stable upper body posture (≤11?mm, ≤2°, p?≤?.01). The results of this study suggest that the inverted chair supports active sitting with backrest support during computer-based office work. However, according to comfort and activity ratings, results should be verified in a future field study with 24 participants.

Practitioner Summary: This experimental laboratory study analyses the feasibility of active sitting with a backrest support during common office work on a new type of dynamic office chair. The results demonstrate that active sitting with a backrest support is feasible on the new but limited on existing chairs.     

一种全局LBF主动轮廓模型

LBF模型的能量函数对于水平集函数是非凸的,从而导致应用LBF模型分割的最终结果对水平集函数的初始化非常敏感。通过凸化LBF模型的能量函数,提出一种全局的LBF模型(GLBF)。该模型针对水平集函数是凸的,从而可以通过任意初始化水平集函数得到全局最优解。此外,该模型不必重新初始化水平集函数为符号距离函数,从而极大地提高运算效率。对灰度不均匀医学图像的分割结果表明,GLBF模型对水平集函数的初始化不敏感,优于传统的LBF模型以及目前具有代表性的LIF模型。     

基于集成的非均衡数据分类主动学习算法

当前,处理类别非均衡数据采用的主要方法之一就是预处理,将数据均衡化之后采取传统的方法加以训练.预处理的方法主要有过取样和欠取样,然而过取样和欠取样都有自己的不足,提出拆分提升主动学习算法SBAL( Split-Boost Active Learning),该算法将大类样本集根据非均衡比例分成多个子集,子集与小类样本集合并,对其采用AdaBoost算法训练子分类器,然后集成一个总分类器,并基于QBC( Query-by-committee)主动学习算法主动选取有效样本进行训练,基本避免了由于增加样本或者减少样本所带来的不足.实验表明,提出的算法对于非均衡数据具有更高的分类精度.     

Investigation of different load changes in wind farm by using FACTS devices

In this study, voltage, active power and reactive power variation of load bus were examined by Flexible AC Transmission Systems devices (FACTS) in grid-connected wind farm. In the system, FACTS devices, Static Synchronous Compensators (STATCOM), Static VAr Compensators (SVC), Static Synchronous Series Compensators (SSSC) and Thyristor Controlled Series Compensators (TCSC), were used. Also, dual feed induction generator (DFIG) was used as generator in the wind farm. This simulation study was done via Matlab/Simulink environment. According to different values of active and reactive loads at different times being attached to the system, shapes of load bus voltage, active power and reactive power variation were given, also the results were presented in tables. In the conclusion of this study, by means of FACTS devices that is used in the wind farm, instability was healed within a short time.     

面向IPv6网络的入侵检测系统的设计与实现

为解决IPv6网络中网络入侵的安全问题,以Snort入侵检测系统为基础,引入了新应用程序接口(NAPI)技术与内存映射技术,同时提出了以用户行为模式为核心的主动防御策略。试验结果表明,该入侵检测系统能有效地提高数据包捕获效率,全面增强IPv6网络检测未知病毒的能力,是一种新型、高效、安全的IPv6网络入侵检测系统。     

三相四开关并联型APF的SVPWM调制算法研究

对三相四开关并联型APF基于空间矢量脉宽调制（SVPWM）技术的指令电流调制算法进行了研究。首先分析了SVPWM调制算法的一般原理,然后在三相六开关SVPWM调制算法的基础上,提出了一种适用于三相四开关并联型APF的SVPWM调制算法,分析了三相四开关并联型APF的SVPWM调制算法与三相六开关之间的区别,利用MATLAB对其进行建模和仿真,验证了该SVPWM调制算法的可行性与有效性。     

基于先验形状信息的水平集图像分割

针对现有水平集方法对于具有强噪声或弱边界的目标进行分割时存在的问题,提出了一种基于形状先验的图像分割方法.该模型采用变分水平集方法,融合了区域特征和边界轮廓特征,并通过相似性匹配选择最佳先验形状.该模型不仅对具有强噪声和弱边界的复杂图像具有较好的分割效果,而且有效地解决了曲线演化的初始轮廓的确定问题.与传统方法进行对比实验,结果表明,该方法具有较好的分割效果和较高的准确率.