基于P2P结构的kNN查询框架  

A Framework for Monitoring Continues k-Nearest Neighbor Queries Based on P2P Systerm

作　　者：宋晓宇[1] 王睿[1] 孙焕良[1] 

机构地区：[1]沈阳建筑大学信息与控制工程学院,辽宁沈阳110168

出　　处：《沈阳建筑大学学报（自然科学版）》2007年第6期1040-1043,共4页Journal of Shenyang Jianzhu University：Natural Science

基　　金：国家自然科学基金项目(60573090);辽宁省自然科学基金(20052006);辽宁省教育厅攻关计划(05L354)

摘　　要：目的提出一种基于P2P结构的移动对象kNN查询框架.方法假设移动对象具有计算处理能力和储存空间,在此基础上进行P2P结构的设计.查询的发起、响应以及计算等不借助中心服务器,而将其转移到移动对象上进行处理.采用距离-响应时间计算算法和kNN查询分段排序算法来进行移动对象kNN查询.结果距离-响应时间计算算法和kNN查询分段排序算法能较好利用框架中各个移动对象响应查询时存在间隔这一特征进行简化计算,提高了框架的工作效率,增强了查询实时性.在更新的过程中通过使用安全时间的概念来减少盲目的对象更新,减少了数据冗余.结论在模拟实验中,框架能充分发挥P2P结构的优点进行移动对象kNN查询,查询响应迅速,查询结果具有较好实时性.This paper proposes a framework for monitoring continues k - nearest neighbor （kNN） queries. Based on the peer- to- peer （P2P） systerm, we suppose that all the moving objects in the framework are in possession of calculating capabilities and memories so that many tasks which used to be possessed by center server are done by moving objects now. In order to fit the capability of moving objects we propose a distance - time computing algorithm and a subsection compose algorithm. Depending on the different distances between the objects and the query point, there are time slots between different moving objects. The algorithms make full use of this characters to improve the efficiency. We adopt safe time algorithm to reduce the unnecessary update. In the simulation, the framework takes full advantage of peer-to-peer configuration, and makes queries efficiently and accurately.

关 键 词：P2P 移动对象kNN查询 距离-响应时间计算算法 kNN查询分段排序算法 

分 类 号：P311.131[天文地球—固体地球物理学]