Adaptive prefetching and storage reorganization in a log-structuredstorage system

We present a storage management system that has the ability to adapt to the data access characteristics of the application that uses it based on collection and analysis of runtime statistics. This feature is especially useful in the storage management layer of database systems, where applications exhibit relatively predictable access patterns. Adaptive reorganization is performed by the storage management system in a manner that optimizes the access patterns of the system for which it is used. We enhance the log-structured storage system that naturally caters for write optimization, with the addition of a statistics collection mechanism to determine data access patterns of applications. The storage system can serve as a testbed for a variety of statistics analysis and clustering mechanisms. Higher level application-specific data clustering mechanisms can be used to override the storage system's low-level clustering mechanisms. In addition, the analysis techniques and reorganization scheme can be used in other storage systems. Performance results from our prototype show potential response time speedups of up to 83 percent over the basic log-structured file system in the best case, using a combination of storage reorganization and prefetching     

基于子带可控响应功率的多声源定位方法

为了提高多个说话人情况下麦克风阵列的定位性能,提出基于子带可控响应功率的多声源定位算法。该算法将语音信号频域分为7个子带,在每个子带计算相位变换加权的可控响应功率函数,在声源空间搜索其最大值得到声源位置的初始估计。根据语音信号频率的稀疏性,这些初始估计包含多个声源的位置,运用会聚聚类算法得到最终的声源位置估计。仿真和实验表明,在有2个说话人,10 dB信噪比,较强混响的条件下,该算法比传统算法的定位正确率提高了约4%,额外率降低了约7%。     

一种基于文本分类的知识树自动构建方法*

针对当前知识管理系统中知识树的创建和维护问题,设计了一种新的基于文本聚类的知识树构建方法。由于从传统的K-means和SOM等文本聚类的结果中难以提取知识树中节点对应的概念和词汇列表,选取PLSA方法进行聚类和知识层次树构建。实验表明,新方法除了在聚类精确度上优于传统方法,聚类结果还包含文档的主题与词汇之间的概率关系,因此新方法在聚类的同时,可以方便地提取知识树上每个节点对应的概念或概念集合。     

A distributed parallel alarm management strategy for alarm reduction in chemical plants

A distributed parallel alarm management strategy based on massive historical alarms and distributed clustering algorithm is proposed to reduce the number of alarms presented to operators in modern chemical plants. Due to the large and growing scale of historical alarms as the basis of analysis, it is difficult for traditional alarm management strategy to store and analyze all alarms efficiently. In this paper, by designing the row key and storage structure in a distributed extensible NoSQL database, the strategy spreads alarm data in a group of commercial machines, which ensures the capacity and scalability of the whole system. Meanwhile, Distributed Parallel Query Model (DPQM) proposed as a unified query model provides efficient query and better integration of distributed platform. Based on the characteristics of alarms and time-delay correlation of alarm occurrence, alarm similarity criteria are proposed to effectively identify repetitive and homologous alarms. In order to group massive alarm data, a new distributed clustering algorithm is designed to work concurrently in MapReduce frameworks. The test results using alarm data from real chemical plants show that the strategy is better than traditional method based on MySQL at system performance, and provides excellent redundant alarm suppression in both normal situation and alarm flooding situation.     

Quantifying bare soil evaporation and its relationship with groundwater depth

The spatial variation of evapotranspiration (ET) and its relationship with vegetation and groundwater depth in the arid Yinchuan plain of China was studied by combining remote-sensing data with groundwater depth observation. The vegetation cover was quantified with the normalized difference vegetation index (NDVI), ET was estimated with the surface energy balance system, and the groundwater depths were measured in monitoring wells. The results show that an area of the plain is covered by vegetation when the NDVI of the area is larger than 0.2 and that groundwater depth is an important factor influencing the ET in the arid region: the deeper the groundwater, the less the ET. The threshold depth of groundwater evaporation is 4 m in the rainy season. A method of quantifying soil evaporation at a regional scale was also presented in this study. The results obtained in this study are useful for future study of the regional water cycle and climate.     

近邻自适应局部尺度的谱聚类算法

提出一种在LUV空间中基于多层次化结构Nystrm方法的自适应谱聚类算法。首先引入LUV色彩空间,避免了RGB色彩空间中色彩辨别阈对分割的影响,在纹理、边缘区域取得了更好的分割效果;其次将谱聚类算法中基于多层次化结构的方法和基于Nystrm采样的方法结合起来,有效减少了运算时间、解决了数据量较大时计算过程中内存溢出的问题;最后在K均值聚类中通过对特征间隙(eigengap)的分析,自适应地选择K值的大小,解决了自动确定聚类数目的问题。将提出的方法在LUV色彩空间中和RGB色彩空间中分别进行图像分割实验,结果表明在LUV色彩空间中取得效果更加理想。同时也将提出的算法与基于Nystrm方法的谱聚类算法(spectral clustering-Nystrm,SC-N)进行比较。实验结果表明,该算法在数据运算量、运行时间和分割结果上都优于SC-N方法。     

集群技术在税务信息集中的应用

基于集群计算思想建立的集群系统可以为市级税务部门信息处理大集中提供服务器平台。集群系统由处理计算节点集群、SAN存储系统、数据库管理集群、集群内部互连网络、外部服务网络和集群管理控制台等6部分组成。通过对山东地税系统使用基于集群技术的税务大集中应用案例的分析和测试，显示了集群技术应用在税务大集中项目中的先进性、优越性和实用性，完全可以满足市级税务大集中的应用需求，具有广阔的应用前景。     

一种基于粒度的粗糙聚类分析方法

基于粗糙集理论,分析等价关系与粒度之间的关系,提出一种基于粒度的粗糙聚类方法。该方法根据数据对象之间的相对相似性形成初始等价关系和等价类,每个等价类对应一个粒度。引入等价关系隶属度因子 ,用于度量等价关系间隶属关系,作为聚类过程一个有效参数,控制聚类的规模。通过迭代计算聚类的有效性,得到优化的聚类结果。聚类过程表明,聚类分析在一个统一的粒度下进行,在样本点之间定义一种等价关系。实验结果证实了该方法的有效性,用规则集描述的聚类结果具有可解释性和合理性。     

基于半监督的超像素谱聚类彩色图像分割算法

近年来谱聚类算法被广泛应用于图像分割领域,而相似性矩阵的构造是谱聚类算法的关键步骤。 针对传统谱聚类算法计算复杂度高难以应用到大规模图像分割处理的问题,提出了基于半监督的超像素谱聚类彩色图像分割算法。该算法利用超像素将彩色图像进行预分割,利用用户提供的少量标记信息构造预分割区域的基于半监督的模糊相似性测度,利用该相似性测度构造预分隔区域的相似性矩阵并通过规范切图谱划分准则对预分割区域进行划分得到最终的图像分割结果。由于少量标记信息和模糊理论的引入,提高了传统谱聚类的分割性能,对比实验也表明该算法在分割效果和计算复杂度上都有较大的改善。     

Smoothing dissimilarities to cluster binary data

Cluster analysis attempts to group data objects into homogeneous clusters on the basis of the pairwise dissimilarities among the objects. When the data contain noise, we might consider performing a smoothing operation, either on the data themselves or on the dissimilarities, before implementing the clustering algorithm. Possible benefits to such pre-smoothing are discussed in the context of binary data. We suggest a method for cluster analysis of binary data based on “smoothed” dissimilarities. The smoothing method presented borrows ideas from shrinkage estimation of cell probabilities. Some simulation results are given showing that improvement in the accuracy of the clustering result is obtained via smoothing, especially in the case in which the observed data contain substantial noise. The method is illustrated with an example involving binary test item response data.     

Linking groundwater simulation and reservoir system analysis models: The case for California’s Central Valley

Groundwater is an important resource. In many developed basins it meets part or all of the water demands. In addition, the management of groundwater resources directly impacts stream flows through stream-aquifer interactions. Yet many reservoir system analysis models that are used for the management of surface water resources either include a simplified representation of the groundwater flow dynamics or rely on surrogate models (linear response functions, artificial neural networks, etc.) which are trained using more complex groundwater models. These approaches may introduce restrictive, sometimes inaccurate, representation of the groundwater flow dynamics and additional modeling steps. In this study a reservoir system analysis model that utilizes an LP solver is linked directly to a non-linear, three-dimensional, finite element groundwater model. The linked model is a general-purpose model and can be applied to any basin. Some of the features of the linked model are showcased by an application to California's Central Valley.     

基于灰色聚类的汽车健康状态评估

文中基于汽车健康管理的概念提出对汽车健康状态进行评估。汽车健康状态评估作为健康管理的重要内容具有重要的现实研究意义。汽车健康状态评估的结果可以为车主提供维修提醒,为汽车维修人员提供初步维修决策,其满足汽车健康管理的需求。汽车系统是一个灰色系统,文中采用灰色理论与聚类分析相结合的灰色聚类方法进行汽车健康状态评估,根据灰色聚类决策方法的原理,分析了汽车健康状态评估指标体系,然后使用模糊层次分析法确定各指标权重,最后利用灰色聚类方法对汽车健康状态等级进行评估。实例表明,基于灰色聚类理论建立的汽车健康状态评估模型科学合理、方便快捷,并具有很好的实用性。     

一种新型基于GPS的无线自组网分群算法

设计了一种新的基于GPS的Ad Hoc网络分群算法,在群初始化充分利用节点的地理坐标进行分群,使得分群更加合理;在群的维护管理方面,引入了候选群首辅助管理分群,使Ad Hoc网络的分群保持更加稳定可靠。基于该种分群算法设计的Ad Hoc网络移动管理系统,能够实时跟踪、显示移动节点在地图上的位置信息,解决了Ad Hoc网络中由于节点移动所造成的移动管理等问题。     

基于子系统间物流联接的大系统柔性分析

研究了对带有子系统间物流联接的大系统的柔性分析方法，采用化工过程系统中的柔性指数来衡量大系统的柔性大小。在化工过程系统柔性指数的计算方法的基础上，本文首先提出增加三组附加约束来提高计算大系统的柔性指数的速度；然后，本文又提出了逐步计算出每个子系统的柔性指数的方法，从而可以更全面地了解大系统的柔性状况；最后。用一个海水利用产业链作为案例研究表明了上述方法的有效性。本文的方法在该类大系统，例如生态工业系统的建设中具有重要的作用。     

基于案例推理的信息系统业务流程知识重用技术研究

本文在研究信息系统业务流程设计的知识表示与重用技术的基础上,提出了一种基于案例的智能化设计知识重用方法进行信息系统业务流程设计.重用方法包含案例表示、基于相似度算法和智能聚类的案例检索、案例优化、案例库组织四个方面.综合四个方面,实现了基于案例推理的业务流程案例检索和知识重用实验系统原型,并进行了实验.实验结果表明本重用方法具有较好的可行性、有效性、效率与质量.     

灰色聚类法在软件项目风险评价中的应用

软件风险的控制在当今软件开发过程中显得越来越重要,而软件项目风险评价的好坏直接影响到风险的管理。目前,大多数软件项目风险评估技术本质上都是基于主观的,而客观地依赖于度量技术进行软件项目风险分析的研究还很少。针对软件项目风险管理的要求,建立了评价指标体系,提出了基于灰色聚类的软件项目风险评价方法。文中案例分析证明了该方法的科学性、客观性与合理性。     

Ex-ante assessment of costs for reducing nitrate leaching from agriculture-dominated regions

In sandy regions of the Netherlands, current agricultural practices result in high nitrate concentrations in the upper groundwater, endangering water resources. A methodology to estimate costs of region specific solutions is applied to a groundwater protection area. It requires that for existing farms in the region variants of management are described that differ in nitrogen surplus. Costs of variants relate to investments and losses in income due to changes in farm management additional to those required by existing legislation. The expected nitrate concentration for each combination of variant and groundwater class is calculated from nitrogen and precipitation surplus and a leaching fraction. A linear programming model indicates the optimal spatial allocation of variants such that desired regional and sub-regional nitrate concentrations are reached at minimum regional costs. In the case study, current land use is grouped into non-dairy farms and various types of dairy farms and non-agricultural land use. Management variants are described for most of these groups. The LP model was used to analyze the effect of the following conditions: (a) the scale at which the tolerated maximum NO₃ concentration is effectuated; (b) the level of this maximum; (c) the hydrological situation; (d) the way farms are grouped; and (e) the biophysical situation of farms. Total regional costs show a negative, but not necessarily linear, relation with the resulting regional nitrate concentration. This relation is strongly influenced by the actual situation regarding management at the farms and the groundwater table depths in the region. For each region a specific cost–benefit relation might apply. Higher groundwater tables result in strongly reduced costs for required changes in farm management. Depending on costs and physical feasibility, measures that result in a higher groundwater table could therefore be used in addition to or replacement of measures at farm level. Calculated costs are influenced by the method of grouping of farms into farm types, the largest effect due to differentiating intensities of production. Differences between farms need therefore be considered when analyzing possibilities for and costs of measures to reduce nitrate leaching. Political decisions, specifically about the scale at which the maximum nitrate level should be effectuated (farm-type or region) also have a strong effect on the cost–benefit relation. Requiring all farm types to comply with a given maximum nitrate concentration, resulting in a specific regional average, is more costly than setting a similar regional maximum and allowing regional averaging. Methodology and preliminary results were discussed with policy-makers and representatives of farmers, environmentalists and groundwater extracting companies. They felt that these discussions helped focussing the interaction between the various groups of stakeholders. This is due to the fact that assumptions about effects of changes in land use and costs involved are made explicit and to the possibility of a more objective weighing of interests since wishes and conditions are translated into goals and constraints of the linear programming model.     

基于IG-NRS与ICK双向压缩的KMS自学习案例知识匹配

案例知识匹配可有效缓解知识过载问题,确保知识应用水平。针对知识管理系统自学习案例知识匹配的冗余性问题,提出了一种基于IG-NRS与ICK的双向压缩方法。该方法首先设计NRS的改进模型IG-NRS,据此约简案例知识属性集,实现邻域决策系统的纵向压缩;在此基础上,通过引入谱聚类判别并剔除不一致案例知识实现其横向压缩;再藉由知识视图相似度锁定目标案例知识簇与最相似案例知识,从而确定知识匹配结果。在多个UCI数据集上的实验结果表明,该方法能有效减少知识管理系统自学习案例知识的冗余,取得更高的知识匹配效率和有效性。     

一种基于人工免疫系统的聚类算法

该文根据数据分析中聚类判断所遵循的原则,模拟抗体捕获抗原的机制,提出了一种基于人工免疫系统的聚类算法。该方法最终可以获得全局最优解,并且具有本质上的并行性、计算效率高、聚类能力强等优点。