A fuzzy clustering-based hybrid method for a multi-facility location problem

A fuzzy clustering-based hybrid method for a multi-facility location problem is presented in this study. It is assumed that capacity of each facility is unlimited. The method uses different approaches sequentially. Initially, customers are grouped by spherical and elliptical fuzzy cluster analysis methods in respect to their geographical locations. Different numbers of clusters are experimented. Then facilities are located at the proposed cluster centers. Finally each cluster is solved as a single facility location problem. The center of gravity method, which optimizes transportation costs is employed to fine tune the facility location. In order to compare logistical performance of the method, a real world data is gathered. Results of existing and proposed locations are reported.     

K-means method for rough classification of R&D employees' performance evaluation

This paper proposes an approach that can roughly cluster a data set with fuzzy linguistic entries as a prior data arrangement for performance evaluation of R&D employees. The extension principles of fuzzy linguistic numbers are used to modify the K‐means method for handling the linguistic data set. We define the absolute difference of fuzzy linguistic variables as their fuzzy distance. Based on this definition, the K‐means approach can be modified slightly for clustering purposes. The performance of employees engaged in designing and R&D‐oriented jobs is possibly related to some qualitative attributes and the evaluation of such attributes for each employee has a tendency toward semantic scales. In the proposed approach, the supervisor can evaluate the performance of each employee directly with a semantic scale. The modified K‐means approach can roughly cluster their performance into different classes in advance of applying some other sophisticated processes.     

Self-Organizing Tunnel Peers

Tunneling is an important approach in IPv6 transition techniques. The tunnel broker model provides a way to build virtual IPv6 networks without manual configuration.However, neither it adapts performance variation on the IPv4 infrastructure,nor it is a scalable solution for a wide-area IPv6 networking environment. In this paper, a self-organizing tunnel peer (SOTP)model is presented. Tunnel peers are clustered in the SOTP system so that optimization is scalable. Four primitive operations related to cluster construction - arrest,release,division and death - endow the system with the nature of self-organization.Occurrence and behavior of the operations are decided by criteria on the IPv4 end-to-end performance; hence measurement is an indispensable component of the system. The metabolism of cluster relaxes the requirement to accuracy of measurement and optimization.     

An improvement of cluster variation method entropy functional for bcc alloys

The entropy functional in the tetrahedron approximation of the cluster variation method (CVM) for the body-centered cubic structure has been modified to yield accurate consolute temperature for a phase separating system. The improvement is achieved by a change in the multiplicity of the basic tetrahedron cluster and approximately accounts for truncation errors in the entropy functional of CVM owing to finite size of the basic cluster. We demonstrate that the phase diagram, thermodynamic properties and short range order parameters show improved agreement with more accurate results, without involving additional computational burden. Hence, the modified entropy functional could be used as a standard for CVM.     

Comparing the performance of traditional cluster analysis, self-organizing maps and fuzzy C-means method for strategic grouping

Strategic group analysis comprises of clustering of firms within an industry according to their similarities with respect to a set of strategic dimensions and investigating the performance implications of strategic group membership. One of the challenges of strategic group analysis is the selection of the clustering method. In this study, the results of the strategic group analysis of Turkish contractors are presented to compare the performances of traditional cluster analysis techniques, self-organizing maps (SOM) and fuzzy C-means method (FCM) for strategic grouping. Findings reveal that traditional cluster analysis methods cannot disclose the overlapping strategic group structure and position of companies within the same strategic group. It is concluded that SOM and FCM can reveal the typology of the strategic groups better than traditional cluster analysis and they are more likely to provide useful information about the real strategic group structure.     

聚类算法在供水管网节点选择中的应用

在开发供水管网调度决策系统时,需对供水管网数百个节点的压力变化及相互间关系进行分析,工作量较大。针对此问题,采用系统聚类中的类平均法将节点聚类,并利用离均差从各类中筛选出代表点。提出针对数据采集时的缺失值及节点间采集时刻不一致问题的数据预处理方法。经实例检验,得出的节点具有良好的代表性及全面性,可满足工程需求。     

Support vector description of clusters for content-based image annotation

Continual progress in the fields of computer vision and machine learning has provided opportunities to develop automatic tools for tagging images; this facilitates searching and retrieving. However, due to the complexity of real-world image systems, effective and efficient image annotation is still a challenging problem. In this paper, we present an annotation technique based on the use of image content and word correlations. Clusters of images with manually tagged words are used as training instances. Images within each cluster are modeled using a kernel method, in which the image vectors are mapped to a higher-dimensional space and the vectors identified as support vectors are used to describe the cluster. To measure the extent of the association between an image and a model described by support vectors, the distance from the image to the model is computed. A closer distance indicates a stronger association. Moreover, word-to-word correlations are also considered in the annotation framework. To tag an image, the system predicts the annotation words by using the distances from the image to the models and the word-to-word correlations in a unified probabilistic framework. Simulated experiments were conducted on three benchmark image data sets. The results demonstrate the performance of the proposed technique, and compare it to the performance of other recently reported techniques.     

Linux集群实时监控系统的一种实现方法

随着云计算技术的日益发展, Linux集群以造价低廉、易于扩充等优势得到了愈来愈广泛的应用. 为了更好地发挥集群性能, 充分利用集群节点的资源, 对集群性能进行实时监控是很有必要的. 提出了一种Linux集群监控器设计与实现方法. 该方法通过每隔一段时间采集节点机/proc虚拟文件系统中的信息, 如CPU和内存使用情况等. 经过过滤后, 通过socket传输给监控服务器. 论文首先给出了监控器的总体设计方案, 整个监控系统由守护在管理节点上的信息管理服务器进程和运行在各个计算节点上的采集器进程组成. 然后分采集器和信息管理器两大部分, 分别介绍了其具体的设计框架和其采用的关键技术. 采集器分主要由信息采集、信息处理和信息传送3 个模块组成, 分别采用3 个线程来完成. 信息管理器采用了线程池技术, 用以接受采集器发送过来的传输请求. 实践证明, 该系统可以很好地满足实时监控Linux 集群性能的需要.     

Stability of a hierarchical clustering

Clustering algorithms have the annoying habit of finding clusters even when the data are generated randomly. Verifying that potential clusterings are real in some objective sense is receiving more attention as the number of new clustering algorithms and their applications grow. We consider one aspect of this question and study the stability of a hierarchical structure with a variation on a measure of stability proposed in the literature.^(1,2)Our measure of stability is appropriate for proximity matrices whose entries are on an ordinal scale. We randomly split the data set, cluster the two halves, and compare the two hierarchical clusterings with the clustering achieved on the entire data set. Two stability statistics, based on the Goodman-Kruskal rank correlation coefficient, are defined. The distributions of these statistics are estimated with Monte Carlo techniques for two clustering methods (single-link and complete-link) and under two conditions (randomly selected proximity matrices and proximity matrices with good hierarchical structure). The stability measures are applied to some real data sets.     

Newtonian clustering: An approach based on molecular dynamics and global optimization

Given a data set, a dynamical procedure is applied to the data points in order to shrink and separate, possibly overlapping clusters. Namely, Newton's equations of motion are employed to concentrate the data points around their cluster centers, using an attractive potential, constructed specially for this purpose. During this process, important information is gathered concerning the spread of each cluster. In succession this information is used to create an objective function that maps each cluster to a local maximum. Global optimization is then used to retrieve the positions of the maxima that correspond to the locations of the cluster centers. Further refinement is achieved by applying the EM-algorithm to a Gaussian mixture model whose construction and initialization is based on the acquired information. To assess the effectiveness of our method, we have conducted experiments on a plethora of benchmark data sets. In addition we have compared its performance against four clustering techniques that are well established in the literature.     

Dynamic local search based immune automatic clustering algorithm and its applications

Based on clonal selection mechanism in immune system, a dynamic local search based immune automatic clustering algorithm (DLSIAC) is proposed to automatically evolve the number of clusters as well as a proper partition of datasets. The real based antibody encoding consists of the activation thresholds and the clustering centers. Then based on the special structures of chromosomes, a particular dynamic local search scheme is proposed to exploit the neighborhood of each antibody as much as possible so to realize automatic variation of the antibody length during evolution. The dynamic local search scheme includes four basic operations, namely, the external cluster swapping, the internal cluster swapping, the cluster addition and the cluster decrease. Moreover, a neighborhood structure based clonal mutation is adopted to further improve the performance of the algorithm. The proposed algorithm has been extensively compared with five state-of-the-art automatic clustering techniques over a suit of datasets. Experimental results indicate that the DLSIAC is superior to other five clustering algorithms on the optimum number of clusters found and the clustering accuracy. In addition, DLSIAC is applied to a real problem, namely image segmentation, with a good performance.     

基于节点能力的Hadoop集群任务自适应调度方法

针对当前Hadoop集群固有的任务级调度分配方法在运行中存在的负载分布不均的现象,着重对集群节点的执行能力进行了分析与研究.提出了一种基于节点能力的任务自适应调度分配方法.该方法根据节点历史和当前的负载状态,以节点性能、任务特征、节点失效率等作为节点任务量调度分配的依据,并使各节点能自适应地对运行的任务量进行调整.实验结果表明集群的总任务完成时间明显地缩减,各节点的负载更加均衡,节点资源的利用更为合理.     

非平衡进程到达模式下MPI广播的性能优化方法

为了提高非平衡进程到达(unbalanced process arrival,简称UPA)模式下MPI广播的性能,对UPA模式下的广播问题进行了理论分析,证明了在多核集群环境中通过节点内多个MPI进程的竞争可以有效减少UPA对MPI广播性能的影响,并在此基础上提出了一种新的优化方法,即竞争式流水化方法(competitive and pipelined method,简称CP).CP方法通过一种节点内进程竞争机制在广播过程中尽早启动节点间通信,经该方法优化的广播算法利用共享内存在节点内通信,利用由竞争机制产生的引导进程执行原算法在节点间通信.并且,该方法使节点间通信和节点内通信以流水方式重叠执行,能够有效利用集群系统各节点的多核优势,减少了MPI广播受UPA的影响,提高了性能.为了验证CP方法的有效性,基于此方法优化了3种典型的MPI广播算法,分别适用于不同消息长度的广播.在真实系统中,通过微基准测试和两个实际的应用程序对CP广播进行了性能评价,结果表明,该方法能够有效地提高传统广播算法在UPA模式下的性能.在应用程序的负载测试实验结果中,CP广播的性能较流水化广播的性能提高约16%,较MVAPICH21.2中广播的性能提高18%～24%.     

A novel soft computing framework for solar radiation forecasting

Accurate forecasting of renewable-energy sources plays a key role in their integration into the grid. This paper proposes a novel soft computing framework using a modified clustering technique, an innovative hourly time-series classification method, a new cluster selection algorithm and a multilayer perceptron neural network (MLPNN) to increase the solar radiation forecasting accuracy. The proposed clustering method is an improved version of K-means algorithm that provides more reliable results than the K-means algorithm. The time series classification method is specifically designed for solar data to better characterize its irregularities and variations. Several different solar radiation datasets for different states of U.S. are used to evaluate the performance of the proposed forecasting model. The proposed forecasting method is also compared with the existing state-of-the-art techniques. The comparison results show the higher accuracy performance of the proposed model.     

An enhanced classification method comprising a genetic algorithm, rough set theory and a modified PBMF-index function

This study proposes a method, designated as the GRP-index method, for the classification of continuous value datasets in which the instances do not provide any class information and may be imprecise and uncertain. The proposed method discretizes the values of the individual attributes within the dataset and achieves both the optimal number of clusters and the optimal classification accuracy. The proposed method consists of a genetic algorithm (GA) and an FRP-index method. In the FRP-index method, the conditional and decision attribute values of the instances in the dataset are fuzzified and discretized using the Fuzzy C-means (FCM) method in accordance with the cluster vectors given by the GA specifying the number of clusters per attribute. Rough set (RS) theory is then applied to determine the lower and upper approximate sets associated with each cluster of the decision attribute. The accuracy of approximation of each cluster of the decision attribute is then computed as the cardinality ratio of the lower approximate sets to the upper approximate sets. Finally, the centroids of the lower approximate sets associated with each cluster of the decision attribute are determined by computing the mean conditional and decision attribute values of all the instances within the corresponding sets. The cluster centroids and accuracy of approximation are then processed by a modified form of the PBMF-index function, designated as the RP-index function, in order to determine the optimality of the discretization/classification results. In the event that the termination criteria are not satisfied, the GA modifies the initial population of cluster vectors and the FCM, RS and RP-index function procedures are repeated. The entire process is repeated iteratively until the termination criteria are satisfied. The maximum value of the RP cluster validity index is then identified, and the corresponding cluster vector is taken as the optimal classification result. The validity of the proposed approach is confirmed by cross validation, and by comparing the classification results obtained for a typical stock market dataset with those obtained by non-supervised and pseudo-supervised classification methods. The results show that the proposed GRP-index method not only has a better discretization performance than the considered methods, but also achieves a better accuracy of approximation, and therefore provides a more reliable basis for the extraction of decision-making rules.     

Multiple background models for speaker verification using the concept of vocal tract length and MLLR super-vector

In this paper, we investigate the use of Multiple Background Models (M-BMs) in Speaker Verification (SV). We cluster the speakers using either their Vocal Tract Lengths (VTLs) or by using their speaker specific Maximum Likelihood Linear Regression (MLLR) super-vector, and build a separate Background Model (BM) for each such cluster. We show that the use of M-BMs provide improved performance when compared to the use of a single/gender wise Universal Background Model (UBM). While the computational complexity during test remains same for both M-BMs and UBM, M-BMs require switching of models depending on the claimant and also score-normalization becomes difficult. To overcome these problems, we propose a novel method which aggregates the information from Multiple Background Models into a single gender independent UBM and is inspired by conventional Feature Mapping (FM) technique. We show that using this approach, we get improvement over the conventional UBM method, and yet this approach also permits easy use of score-normalization techniques. The proposed method provides relative improvement in Equal-Error Rate (EER) by 13.65?% in the case of VTL clustering, and 15.43?% in the case of MLLR super-vector when compared to the conventional single UBM system. When AT-norm score-normalization is used then the proposed method provided a relative improvement in EER of 20.96?% for VTL clustering and 22.48?% for MLLR super-vector based clustering. Furthermore, the proposed method is compared with the gender dependent speaker verification system using Gaussian Mixture Model-Support Vector Machines (GMM-SVM) super-vector linear kernel. The experimental results show that the proposed method perform better than gender dependent speaker verification system.     

Performance of SOR methods on modern vector and scalar processors

The building-cube method (BCM) is a new generation algorithm for CFD simulations. The basic idea of BCM is to simplify the algorithm in all stages of flow computation to achieve large-scale simulations. Calculation of a pressure field using the Successive Over Relaxation (SOR) method consumes most of the total execution time required for BCM. In this paper, effective implementations on modern vector and scalar processors are investigated. NEC SX-9 and Intel Nehalem-EX are the latest vector and scalar processors. Those processors have much higher peak performances than their previous-generation processors. However, their memory bandwidth improvement cannot catch up with the performance improvement of processors. This is the so-called memory wall problem. In our paper, we discuss optimization techniques for implementation of the SOR method based on architectural characteristics of these modern processors, and evaluate their effects on the sustained performances of these processors for BCM.     

An Analysis of Quasi-Monte Carlo Integration Applied to the Transillumination Radiosity Method

This paper presents an enhanced transillumination radiosity method that can provide accurate solutions at relatively low computational cost. The proposed algorithm breaks down the double integral of the gathered power to an area integral that is computed analytically and to a directional integral that is evaluated by quasi-Monte Carlo techniques. Since the analytical integration results in a continuous function of finite variation, the quasi-Monte Carlo integration that follows the analytical integration will be efficient and its error can be bounded by the Koksma-Hlawka inequality. The paper also analyses the requirements of the convergence, presents theoretical error bounds and proposes error reduction techniques. The theoretical bounds are compared with simulation results.     

对抗彩色贴片：一种针对DNNs的对抗攻击手段

深度神经网络（deep neural networks,DNNs）在图像分类、分割、物体检测等计算机视觉应用方面表现出了先进的性能。然而,最近的研究进展表明,DNNs很容易受到输入数据的人工数字扰动（即对抗性攻击）的干扰。深度神经网络的分类准确率受到其训练数据集的数据分布的显著影响,而输入图像的颜色空间受到扭曲或扰动会产生分布不均匀的数据,这使深度神经网络更容易对它们进行错误分类。提出了一种简单且高效的攻击手段——对抗彩色贴片（AdvCS）,利用粒子群优化算法优化彩色贴片的物理参数,实现物理环境下的有效攻击。首先,提出了一个图片背景颜色变化的数据集,通过在ImageNet的一个子集上用27个不同的组合改变他们的RGB通道颜色,研究颜色变化对DNNs性能的影响。在提出的数据集上对几种最先进的DNNs架构进行了实验,结果显示颜色变化和分类准确率损失之间存在显著相关性。此外,基于ResNet 50架构,在提出的数据集上演示了最近提出的鲁棒训练技术和策略（如Augmix、Revisiting、Normalizer Free）的一些性能实验。实验结果表明,这些鲁棒训练技术可以提高深度神经网络对颜色变化的鲁棒性。然后,使用彩色半透明贴片作为物理扰动,利用粒子群优化算法优化其物理参数,将其置于摄像头上执行物理攻击,实验结果验证了提出的方法的有效性。     

K-Means Clustering Versus Validation Measures: A Data-Distribution Perspective

K-means is a well-known and widely used partitional clustering method. While there are considerable research efforts to characterize the key features of the K-means clustering algorithm, further investigation is needed to understand how data distributions can have impact on the performance of K-means clustering. To that end, in this paper, we provide a formal and organized study of the effect of skewed data distributions on K-means clustering. Along this line, we first formally illustrate that K-means tends to produce clusters of relatively uniform size, even if input data have varied “true” cluster sizes. In addition, we show that some clustering validation measures, such as the entropy measure, may not capture this uniform effect and provide misleading information on the clustering performance. Viewed in this light, we provide the coefficient of variation (CV) as a necessary criterion to validate the clustering results. Our findings reveal that K-means tends to produce clusters in which the variations of cluster sizes, as measured by CV, are in a range of about 0.3–1.0. Specifically, for data sets with large variation in “true” cluster sizes (e.g., $ hbox{CV} ≫ 1.0$), K-means reduces variation in resultant cluster sizes to less than 1.0. In contrast, for data sets with small variation in “true” cluster sizes (e.g., $hbox{CV} ≪ 0.3$), K-means increases variation in resultant cluster sizes to greater than 0.3. In other words, for the earlier two cases, K-means produces the clustering results which are away from the “true” cluster distributions.