一种多维向量并行查询算法  

A Multi-dimensional Vector Parallel Search Algorithms

作　　者：黄海[1] 侯颖[1] 朱圣平[1] 

机构地区：[1]解放军信息工程大学,河南郑州450002

出　　处：《微电子学与计算机》2010年第8期17-19,共3页Microelectronics & Computer

基　　金：国家"八六三"计划项目(2008AA011001)

摘　　要：为了解决多维向量数据快速查询问题,在查询范围上限已知的条件下,通过对数据集合采用最近邻准则进行空间划分,构造一种多叉扩展平衡索引树,并设计了索引树的串行和并行查询算法.最后,对并行查询算法的性能进行了分析,测试结果验证了该方法的有效性.To solve the quick search problem of multi-dimensional vector data, this paper constructed a multi-branch extended and balanced index tree, by using the nearest neighbor criterion to realize the space division of the data sets, under the condition that the upper limit of the query scope is known. Then, the serial and parallel search algorithms of the index tree were designed. At last, the capability of the parallel search algorithms was analysed and the test results validated the validity of this method.

关 键 词：多维索引结构 并行计算 相似查询 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]