无线传感器网络分簇路由节能研究

在大规模传感和环境监测中,节约能源延长传感器节点生命已成为无线传感器网络最重要的研究课题之一。提供合理的能源消耗和改善无线网络生命周期的传感器网络系统,必须设计一种新的有效的节能方案和节能路由体系。方案采用一种聚类算法减少无线传感器网络的能量消耗,创建一种cluster-tree分簇路由结构的传感器网络。该方案主要目标是做一个理想的分簇分配,减少传感器节点之间的数据传输距离,降低传感器节点能源消耗,延长寿命。实验结果表明,该方案有效地降低了能源消耗从而延长无线传感器网络生命。     

虚拟环境下Web服务动态负载均衡策略改进

为了提高Web服务集群的伸缩性和自动化能力,从虚拟化和负载均衡两方面研究集群系统,对现有负载采集策略做了改进,设计并实现了一种可根据负载值自动控制集群规模的模型XCluster。新模型运行在Xen提供的虚拟化环境中,实时监视宿主机层和虚拟机层的负载状态,随着集群系统总负载的增长,逐渐引入新的虚拟机来扩大集群规模,同时将任务合理分配到各个虚拟机节点上;当总负载下降时,逐渐关闭虚拟机缩小集群规模,释放出来的硬件资源又可以提供给其他集群系统使用。理论分析和实验结果表明,XCluster只需占用很少的网络通信量完成信息收集和命令下达,能够充分利用虚拟机易于管理的优势完成后端节点的调度,并且在任务总量相同的情况下,使用尽可能少的集群节点来执行任务。     

微电流电解对尿素降解的机理及影响因素研究

采用微电流电解技术对模拟游泳池水中的尿素进行降解处理,探索了电解对尿素的去除机理,研究了电流密度和氯离子浓度等因素对电解去除尿素效果的影响。结果表明:以钌钛和不锈钢分别做阳极和阴极材料,当水溶液中有较高浓度的氯离子存在时,电解过程中产生的活性氯使得尿素被氧化,尿素降解产物中无氨氮、氯胺等二次污染物,尿素极有可能通过电解生成了N2等气体挥发到空气中。电流密度和氯离子浓度对尿素的去除效率有很大影响,随着电流密度和氯离子浓度的增加,尿素去除效率逐渐上升。在500 mg/L的氯离子浓度条件下,电流密度为12 m A/cm2时,电解处理30 min时尿素去除率可达99.5%。将微电流电解技术应用于游泳池水中的尿素去除,无需添加电解质,即能达到较高的尿素去除效果,采用该技术对游泳池水中的尿素进行降解去除完全可行。     

服务器集群技术综述

文章叙述了服务器集群技术的概念,集群技术在电子商务、网络信息系统应用领域中的重要地位,集群技术的分类,几种不同集群技术的优缺点。     

面向多用户的集群资源部署策略

集群系统中面向多用户的资源部署策略优先考虑的是资源能够在不同用户之间进行公平分配,同时实现用户之间的资源共享。为了充分利用集群资源,用户以一定的原则通常是采用权重比的方式共用集群资源。该策略并不关心每个用户所获取的资源总量,只关心用户所获取的资源比例是否总是符合原则,这样可以保证空闲资源总是能够被充分利用。该策略包括资源的定额分配和自主申请两部分,以满足用户的需求。同时提出了采用用户占用资源量与占用时间共同作为公平分配依据的策略,以适应更加复杂的情况。     

Understanding post‐adoption IS usage stages: an empirical assessment of self‐service information systems

Organisations have invested in self‐service information systems (IS) to provide a direct interface for service delivery. Enriching the usage of these systems can provide organisations with immense benefits. However, limited research has been directed towards understanding post‐adoption IS usage behaviour in general and specifically in the context of self‐service IS. This study proposes post‐adoption IS usage behaviour as a broader concept constituting feature level usage of IS, integration of IS in the work system and exploration of new uses of IS. We evaluate how the new conceptualisation can be used to classify users at different stages of self‐service IS usage. Further, we examine user perceptions that differentiate among the users situated at different self‐service IS usage stages. Data were collected in the context of a self‐service Web‐based IS to validate the post‐adoption IS usage constructs and to examine the proposed thesis. The newly developed conceptual structure and measures for post‐adoption IS usage behaviour exhibit strong psychometric properties. The analysis shows three distinct post‐adoption IS usage stages and highlights that usefulness, user‐initiated learning, ease of use, satisfaction and voluntariness of use differentiate users at the different stages of post‐adoption IS usage. The results show that these variables aggregate into value confirmation and learning orientation as two higher‐level concepts. Further, we evaluate the predictive efficacy of the research model in classifying users into different post‐adoption self‐service IS usage stages. Implications are drawn for future research.     

Water quality in the southern Tibetan Plateau: chemical evaluation of the Yarlung Tsangpo (Brahmaputra)

Yarlung Tsangpo (Brahmaputra) is the largest river system draining the northern slopes of the Himalayan ranges on the southern Tibetan Plateau. It remains one of only two large non‐regulated rivers in China. In this paper the chemical composition of Yarlung Tsangpo and its major tributaries (Raga Tsangpo, Nyangchu and Lhasa River) are studied. Water samples (n = 55) were collected and measured for major ions, trace elements and nutrients in order to: (1) define the present chemical quality of this water course; (2) address possible mechanisms governing the water chemical compositions, and (3) identify potential sources for contaminants. Multivariable analysis shows that geology and climate are the major explanatory variables for the spatial variation in water chemistry in this river system. In general, water chemistry is mainly controlled by carbonate weathering, with Ca²⁺ and HCO being the dominant ions. In addition, runoff from brackish/saline lakes and geothermal waters, enriched in Na⁺, Cl^?, SO, Mg²⁺ and Li, are major contributors of elevated concentrations of these solutes in the headwater regions resulting in a relatively high loading of total dissolved solids (TDS, 146–397 mg L^?1). Levels of most heavy metals and total dissolved nutrients were generally found to be low. However, elevated As concentration (avg. 95 μg L^?1) in the headwaters and additions from untreated wastewater were evident at some locations. Copyright © 2009 John Wiley & Sons, Ltd.     

可拓K近邻算法在数据聚类分析中的应用

针对区间值数据的数据聚类问题,根据可拓学关联函数的定义,提出可拓距离的概念来度量数据之间的距离,利用K近邻的思想,根据可拓距离的大小对数据集的目标属性进行投票选择进行分类,设计了可拓K近邻算法（Extension K Nearest Neighbor,EKNN）。最后利用UCI的两个基准数据集Iris植物样本数据和糖尿病数据库PIDD进行验证,首先通过免疫网络约简算法对条件属性进行最小属性约简,然后利用EKNN算法分析和比较不同最小约简属性下的分类准确率。     

实时视频中的车辆运动轨迹的提取和聚类

利用运动物体轨迹的方向性、运动性和相互关系等典型特征对物体的运动轨迹进行聚类。首先利用改进的加权矢量Hausdorff距离作为度量运动物体轨迹相似度的方法,从而使之适用于空间距离有差别的运动物体轨迹的谱聚类问题;然后基于等周分割（ISO）算法,构造轨迹相似度矩阵,完成轨迹的粗聚类;最后利用轨迹的方向性特征和轨迹类间距对轨迹进行二次聚类,得到最终的轨迹聚类结果。     

集群式软件平台敏捷测试

在敏捷开发模式的环境下,对以集群存储业务为导向的集群平台的测试与其他产品的测试有很多不同的地方,需要采用有效的策略和方法。首先介绍了敏捷测试的理论和敏捷测试传统的增量模型,在此基础上结合集群平台敏捷开发中实际的问题提出了适合平台敏捷测试的增量模型和对集群平台用户故事测试时用到的仿真环境的搭建和有效的测试方法。