云环境中改进模糊聚类的资源聚合  

Resources Polymerization Based on Improved Fuzzy Clustering in Cloud Environment

作　　者：王溢琴[1] 秦振吉[1] 

机构地区：[1]晋中学院计算机科学与技术学院,山西晋中030600

出　　处：《计算机仿真》2013年第4期365-368,共4页Computer Simulation

摘　　要：研究云中高性价比的资源聚合问题。针对云中硬件资源聚合时没有考虑到资源的地理位置和运行成本,特别是传统FCM算法聚类时无法比较资源之间位置的关系,体现不出性价比。为解决上述问题,提出了一种改进模糊算法的资源聚合策略。该策略首先定义描述资源位置关系的加权因子并构建新的目标函数;然后在资源属性矩阵中标准化数据,重复迭代改进目标函数优化数据集的划分;最后定义了聚类集群综合性能评价指标,能从若干个子类中选取高性价资源池的目的。仿真结果表明,改进算法和性能评价指标得到的聚类结果较好,适合云环境中资源性价比高的整合。The paper studied of cost - effective resources polymerization problem, and put forward a new re- sources polymerization strategy based on improved fuzzy algorithm. Firstly, the strategy constructs a new objective function based on proposed weighted factor. Secondly, it standardizes data in the resource attribute matrix. An objec- tive function is iterated to optimize the data set. Finally, it defines cluster comprehensive performance evaluation in- dex, which can select the high cost - effective pool from several subclasses. The experimental results show that it has a better clustering results according to the improved algorithm and the performance evaluation indexes, and can be applied to resources integration with high performance.

关 键 词：资源属性 加权因子 目标函数 评价函数 集群综合性能 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]