一种基于信息素的蚁群聚类算法

提出了一种改进的蚁群聚类分析算法,通过改进LF算法中群体相似度函数,加入参数的自适应调整策略,利用短期记忆和网格信息素的局部分布控制蚂蚁的随机移动,并结合蚂蚁速度动态变化、半径递增、强制放下等特性。采用测试数据和不同的算法进行了对比实验分析,仿真实验结果表明,该算法显示出了较高的稳定性和准确率。     

一种基于Web资源相似性的序列模式聚类方法

在Web行为挖掘中,序列模式聚类是一个很重要的课题,其首要问题就是web序列模式间的相似性度量.以往的多数方法都仅仅针对序列本身进行度量,而忽略了系统中资源本身所存在的关联关系以及用户对资源访问的时间因素.针对该问题,提出了一种基于考虑资源相似性的web访问序列模式的相似度量方法,并且考虑了用户访问资源的时间因素.经过检验,证明能够有效真实地反映实际情况.     

一种有效的启发式聚类算法

本文讨论了一种利用确定性退火技术的启发式聚类算法。它把聚类问题看作一物理系统。通过求解一系列随温度变化的自由能函数的全局极小来得到聚类问题的最优解。算例表明，对传统聚类算法无能为力的几种聚类问题，该算法都得到了比较满意的结果。     

一种有效的均值聚类初始化方法

聚类算法是用来提取有用信息的重要技术,k均值聚类算法是其中应用最为普遍的聚类分析算法.然而,这种聚类算法的主要问题是,最终的聚类结果高度依赖于初始聚类中心.标准的k均值聚类算法使用随机初始中心会得到很差的聚类结果.因此,为了克服标准k-均值聚类算法的不足,本文提出一种基于贡献率的方法来优化初始中心的选择,以便得到一个好的聚类结果.将新提出的初始化方法应用到一些知名的数据集,将其与几种传统的初始化算法相比较,证明新提出的初始化方法具有良好的性能.本文所提出的方法不仅容易理解,而且聚类的迭代次数和执行时间也明显下降.本文的初始化方法可以保证得到一个比较好的聚类结果.     

一种基于星座图聚类的MQAM识别方法

星座图是数字调制信号MQAM的重要特征,提出了一种利用星座图减法聚类对MQAM信号进行识别的新算法。该算法通过分析接收端信号的星座图,提取描述星座图特征的参数,与标准星座图进行相关比较,实现了加性高斯白噪声下数字调制信号MQAM的识别。通过实验证明该方法可以有效地对数字调制信号MQAM进行识别,而且该算法具有复杂度低的优点。尤其当数据长度比较短时,有较好的识别效果。     

一种新的并行聚类算法

为了进一步提高计算效率,提出一种面向计算机集群的并行聚类算法.首先合理划分数据块,之后由集群中的节点并行聚类,最后合并聚类结果.这期间采用动态负载平衡和异构的通讯方式.实验证明该并行算法不仅能得到正确的聚类结果,并能够大幅度提高聚类速度.     

一种基于类中心矫正的层次聚类算法

为了克服传统层次聚类算法由于两类合并造成的中心点偏移的严重缺陷,提出了一种基于类中心矫正的层次聚类算法,从而提高了算法的精确度;同时继承了传统层次聚类对初始中心点的无依赖性;经分析,算法对于已知聚类数和未知聚类数两种情况均有着良好的聚类效果.通过标准数据测试,结果表明新算法的聚类性能与层次聚类算法相比有更高的精确度;并且让新算法用于指导图像分割实验,证明了算法的有效性.     

一种基于DFC聚类算法的图像分割方法

图像分割是一种重要的图像技术,在理论研究和实际应用中都得到了人们的广泛重视。图像分割的方法和种类很多,有些分割运算可直接应用于任何图像,而另一些只能适用于特殊类别的图像。目前,图像分割的方法层出不穷。其中,最具代表性的图像分割算法是基于FCM聚类算法的图像分割方法。然而FCM聚类算法从理论上来说存在着聚类数目无法自动确定及运算的开销太大的缺点,因而限制了这种方法的应用。针对其不足,本文将FCM聚类算法引入到图像分割方法中。数值实验结果显示：新方法分割图像的效果是良好的。     

一种基于调和均值的模糊聚类算法

k调和均值算法用数据点与所有聚类中心的距离的调和平均替代了数据点与聚类中心的最小距离，是一种减小初始值影响聚类结果的有效的聚类方法。本文对k调和均值算法进行扩展，考虑到数据点同时对不同聚类的隶属关系，将模糊的概念应用到聚类中，提出了模糊k调和均值-Fuzzv K—Harmonic Means(FKHM)算法。在中心迭代聚类算法的统一框架的基础上，推导出FKHM算法聚类中心的条件概率表达式以及在迭代过程中的数据点加权函数表达式。以划分相似度作为聚类结果的评价准则，实验表明，FKHM算法在聚类对于初值不敏感的同时提高了聚类结果的精确度，达到较好的聚类效果。     

一种改进的模糊聚类算法

ＦＣＭ（Ｆｕｚｚｙ Ｃ－Ｍｅａｎｓ）算法是一种基于目标函数优化的模糊聚类方法,其收敛地于初始条件敏感。与ＨＣＭ（Ｈａｒｄ Ｃ－Ｍｅａｎｓ）算法相比,ＦＣＭ算法的模糊分割矩阵提供的信息更加丰富。本文采用冗余聚类中心初始化,根据模糊分割矩 列和以及实际的要求逐级减少类别数目。实验结果显示改进的算法得到的收敛中心稳定,并且中以融合有关数据分布的先验知识得到所期望的结果。     

UCRP—一种能量有效的WSN非均匀分簇路由协议

提出了一种新的能量有效的非均匀分簇路由协议—UCRP.根据距离基站的远近将网络分为大小不同的簇;簇内数据传输根据簇范围的大小采用单跳或多跳;簇间数据传输构建多跳路由,簇首选择下一跳节点时将能量与最小跳数路由算法相结合;最后在能量有效的前提下对LEACH协议易受到HELLO flooding攻击提出了安全设想.仿真结果表明:该协议能够有效地均衡簇首和全网能耗,延长网络生存时间.     

Energy Efficient Clustering Scheme (EECS) for Wireless Sensor Network with Mobile Sink

The participants in the Wireless Sensor Network (WSN) are highly resource constraint in nature. The clustering approach in the WSN supports a large-scale monitoring with ease to the user. The node near the sink depletes the energy, forming energy holes in the network. The mobility of the sink creates a major challenge in reliable and energy efficient data communication towards the sink. Hence, a new energy efficient routing protocol is needed to serve the use of networks with a mobile sink. The primary objective of the proposed work is to enhance the lifetime of the network and to increase the packet delivered to mobile sink in the network. The residual energy of the node, distance, and the data overhead are taken into account for selection of cluster head in this proposed Energy Efficient Clustering Scheme (EECS). The waiting time of the mobile sink is estimated. Based on the mobility model, the role of the sensor node is realized as finite state machine and the state transition is realized through Markov model. The proposed EECS algorithm is also been compared with Modified-Low Energy Adaptive Clustering Hierarchy (MOD-LEACH) and Gateway-based Energy-Aware multi-hop Routing protocol algorithms (M-GEAR). The proposed EECS algorithm outperforms the MOD-LEACH algorithm by 1.78 times in terms of lifetime and 1.103 times in terms of throughput. The EECS algorithm promotes unequal clustering by avoiding the energy hole and the HOT SPOT issues.     

Energy Efficient Non-uniform Clustering Division Scheme in Wireless Sensor Networks

Wireless sensor networks can be used to monitor the interested region by multi-hop communication. Since sensor nodes are equipped with energy-limited batteries, energy conservation in such networks is of paramount importance in order to prolong the network lifetime. In this paper, considering the constrained radio range of node, we propose an energy efficient clustering division scheme from the viewpoint of energy consumption. The difference between our scheme and previous schemes is that ours is a non-uniform clustering hierarchy. With the algorithm that is proposed by this paper, we can divide the cluster into multiple non-uniform concentric rings and obtain the optimal thickness of each ring. Motivated by the derived results, every sensor node can adjust its radio range for transmission. Our extensive simulation results indicate that the proposed non-uniform clustering division scheme outperforms the conventional uniform clustering division schemes in terms of energy consumption and lifetime. The future research that should be explored is also discussed finally.

Energy Harvesting Based Efficient Routing Scheme for Wireless Sensor Network

Wireless Sensor Network were deployed in a complex environment where the wide range of complex application is mandatory for the services. Such application includes military, agriculture, healthcare, defense, monitoring, surveillance etc. In general sensor nodes were spatially distributed and deployed in remote fashion, usually they are powered up by batteries. These battery powered sensor nodes are pruned to failure due to its power constrained nature. This led many researchers to explore energy efficient context aware routing for Wireless Sensor Networks. Hence a novel energy harvesting based efficient routing scheme is desirable to address the above stated problem. The key idea is to harvest the energy source from the deployed environment. The proposed routing scheme is tested and validated in MATLAB based simulation test bed. The experimental results shows that the proposed routing scheme is robust and meet all the requirements of routing and promising results for energy usage.     

无线传感器网络中分簇算法能量有效性分析

该文针对无线传感器网络是一种能量受限的网络,首先说明根据节点数目以及分布区域特征,确定合理的成簇数目是分簇算法设计的核心;然后给出节点在通信中使用的能量模型,按照分簇的不同层数,分别对其成簇的数目进行理论上的优化分析,最后通过实验的方法验证优化的分簇算法能够满足能量有效性,对均匀节点能耗分布,提高网络生存时间有着重要作用。     

一种基于超节点和能量优先的无线传感器网络的高效查询算法

文中提出一种基于超节点和能量优先的无线传感器网络的高效查询算法.该算法包括传感器节点的层次聚类算法及基于能量代价模型等支撑技术,主要解决了以下两个问题:(1)数据如何从传感器节点传送到汇聚节点;(2)通过对传感器节点进行聚类,形成超节点,使得在查询过程中减少对无关节点的访问.实验表明该算法在提高无线传感器网络查询效率的情况下,延长网络的使用寿命.     

A New Energy Efficient Clustering Protocol for a Novel Concentric Circular Wireless Sensor Network

Wireless Personal Communications - Researchers concentrate on big data. Wireless sensor network is one of the sources of big data. Wireless sensor network has hundreds of sensor nodes with limited...     

A Mathematical Model for Energy Efficient Secured CH Clustering Protocol for Mobile Wireless Sensor Network

Wireless Personal Communications - Mobile Wireless sensor network encloses several minute sensors with miniature energy transceiver for collecting information in a diverse atmosphere. Network with...     

A Sensor Anonymity Enhancement Scheme Based on Pseudonym for Clustered Wireless Sensor Network

Security problem is an important issue for Wireless Sensor Network. The paper focuses on the privacy protection of WSN applications. An anonymity enhancement tactic based on pseudonym mechanism is presented for clustered Wireless Sensor Network, which provides anonymity for both the sensors within a cluster and the cluster head nodes. Simulation experiments are launched through NS2 platform to validate the anonymity performance. The theoretical analysis and empirical study imply that the proposed scheme based on pseudonym can protect the privacies of both the sensor nodes and the cluster head nodes. The work is valuable and the experimental results are convincible.     

无线传感器网络关键节能技术及节能策略研究

无线传感器网络(WSN,Wireless Sensor Network)是一种由传感器、数据处理单元和通信模块集成的微小节点通过自组织方式构成的网络,其节点通常采用无法替换的电池供电,有限的能量使得节能问题成为无线传感器网络的研究重点方向.本文介绍了传感器的能耗情况,结合已有研究重点分析了关键节能技术及节能策略,并指出了研究方向.