基于增量聚类的语义缓存替换策略  

Replacement strategy for semantic caching based on incremental clustering

作　　者：李东[1] 叶友[1] 杨小鹏[2] 刘振宇[2] 

机构地区：[1]华南理工大学计算机科学与工程学院,广州510640 [2]华南理工大学软件学院,广州510006

出　　处：《计算机应用研究》2008年第12期3610-3613,共4页Application Research of Computers

基　　金：广东省自然科学基金资助项目(480B6040550)

摘　　要：为了提高系统性能以及克服移动计算环境自身的通信带宽窄、网络断接等缺陷,采用语义缓存技术来管理相关数据就显得尤为重要。形式化地给出了语义缓存的相关概念和定义,然后重点分析了语义缓存的FAR(furthest away replacement)替换策略并对其进行改进,提出基于增量聚类的DC-FAR替换策略,最后对FAR和DC-FAR替换策略进行实验分析,从而在理论上和实验上证明了DC-FAR替换策略的优越性。To improve system performance and overcome narrow bandwidth and disconnection of mobile computing environment, it is very important to use the semantic caching to manage location dependent data. This paper gave the formal definitions of semantic caching. Then analyzed the FAR（ furthest away replacement） replacement strategy, developed a new replacement strategy named data clustering FAR（DC-FAR） ,which based on the previous query data. At last, compared the FAR and DC-FAR replacement strategy through the experiment. The experiment shows that the DC-FAR was more effective than FAR in the theory and experimentation.

关 键 词：移动计算 语义缓存 替换策略 位置相关 增量聚类 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]