有效支持空间OLAP的空间数据仓库模型研究  

Research on a spatial data warehouse framework for efficient spatial OLAP

作　　者：陈细谦[1] 迟忠先[1] 徐世宏[1] 

机构地区：[1]大连理工大学计算机科学与工程系,辽宁大连116024

出　　处：《大连理工大学学报》2004年第6期901-905,共5页Journal of Dalian University of Technology

摘　　要：为了实现各类空间联机分析,提供更加全面灵活的空间决策支持,提出了一个空间数据立方体与空间索引结构相互协作的空间数据仓库模型,并同时构建了一个与之协作的空间索引结构aR*Btree.aR*Btree基于R*tree和Btree实现,存储了空间维及时间维的聚集信息和层次关系,可以有效支持区域聚集查询.对各种空间联机分析操作进行了分类,分析表明该模型充分利用了传统数据仓库成熟的建模和联机分析技术,结合了空间索引结构在空间层次结构上的灵活特性,可以有效地支持各类空间联机分析.最后比较了aR*Btree与其他索引结构在区域聚集查询方面的性能.实验结果证明,在区域聚集查询中aR*Btree的节点访问次数和查询执行时间均小于现有索引结构.To realize different kinds of spatial OLAP operations and provide spatial decision support more flexibly, a spatial data warehouse framework is studied, in which spatial data cube cooperates with spatial index structure. By incorporating to it, a spatial index structure named aR~*Btree is proposed to support spatial analysis flexibly on complex spatial hierarchies. The aR~*Btree is an improved and conjoint version of the R~*tree and Btree which stores both the hierarchies relations of time and spatial dimension and the aggregate values of them. The framework takes full advantage of traditional data warehousing technology and the favorable characteristics of index structure in hierarchies′ management. By classifying OLAP operations, it shows how these spatial OLAP operations are efficiently supported in the framework. Finally the performance of range aggregate processing is compared with other indices, and the result shows that the node accesses and executive time of aR~*Btree′s range aggregate queries is superior to those of other structures.

关 键 词：聚集查询 空间索引结构 联机分析 OLAP 空间数据仓库 协作 层次关系 区域 支持 模型研究 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论] F270[自动化与计算机技术—计算机科学与技术]