网络环境下基于差值的聚合查询策略

聚合查询(GNN)在网络环境中是一个相对比较新的操作.与传统的只有一个查询结点的任意性,使得GNN查询比只有一个查询点的kNN查询复杂得多.文章提出了一种基于2个查询点的差值聚合查询,基于单查询点的YPK-CNN,利用双曲线性质把查询空间分成几大部分,在此基础上计算出双曲线的渐近线以简化删减条件,直接删减掉不符合条件的搜索空间.实验结果显示,针对这种特定的差值聚合查询,我们所提出的基于双曲线渐近线过滤策略的查询算法要比基本查询算法在查询响应时间和访问结点次数方面更高效.

Difference Group Query Strategy in Network Environment

Group Nearest Neighbor query is a relatively prevalent new application in network environments.Unlike the traditional kNN query which contains one query point in the system,GNN query maintains several query points.The multiple query points and the randomicity of data distribution in net space make it more complicated for a GNN query compared to a kNN query.This paper puts forward a new Difference Group Query based on two query points,the methods in GNN query which define several query points in a MBR are not suitable for the specific Difference Group Query in our paper.In this paper,based on the algorithm in YPK-CNN,we use the definition of hyperbola,divide the query space into several sections to delete the unsuitable space,furthermore,we develop several asymptotes from the definition of hyperbola,simplify the hyperbola-based filter strategy in the search space.In the experiment and analysis section,from the output of our experiment,we prove that the algorithm with the hyperbola-based filter strategy is much better than the basic query algorithm in the respects of query response time and accessed nodes.