基于粗集理论的视图分层更新方法  

Hierarchical Updating Method Based on Rough Sets Theory in View Maintenance

作　　者：李文海[1] 冯玉才[1] 马晓鸣[1] 尹建章[1] 

机构地区：[1]华中科技大学计算机学院数据库与多媒体技术研究所,武汉430074

出　　处：《计算机科学》2008年第1期121-125,共5页Computer Science

基　　金：国家"八六三"高技术研发计划项目(2004AA4Z3020;2005AA4Z3030);科技部电子政务关键技术应用研究项目(2001BA110B01)

摘　　要：为缩短查询响应时间和提高决策支持能力,数据仓库中存储了大量的实化视图,视图更新成为影响系统整体性能的一个重要因素。本文分析了现有方法处理大规模视图更新的不足,提出一种基于粗集约简发现中间视图依赖层次的方法,利用有效代价确定更新次序以达到更新费用最低,随后分析算法的复杂性并通过实验证明其有效性。For the sake of acquiring the quick response and enhancing the power of the decision support, data warehouses maimain large numbers of materialized views. And view updating becomes a significam factor of the overall performance. This paper analyzes the defect of the existing methods in handling the updating of the massive views, and proposes a new method based on rough sets theory to discover the correlative hierarchy of the mutual views. The valid cost is utilized to acquire the up dating order with the minimal expens. The validity and complexity of the algorithm are given in the end.

关 键 词：数据仓库 实化视图 层次更新 粗集 约简 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]