基于遗传算法的分布式数据库查询优化研究  

Research of Query Optimization Based on Genetic Algorithm in Distributed Database

作　　者：帅训波[1] 马书南[2] 周相广[1] 龚安[3] 

机构地区：[1]中国石油勘探开发研究院廊坊分院地球物理与信息研究所,河北廊坊065007 [2]北京工业大学计算机学院,北京100022 [3]中国石油大学计算机与通信工程学院,山东东营257061

出　　处：《小型微型计算机系统》2009年第8期1600-1604,共5页Journal of Chinese Computer Systems

基　　金：中国石油天然气集团科研基金项目(KF10801-1)资助

摘　　要：在收益半连接研究基础之上,本文提出一种新的查询执行计划代价模型,它包括关系副本选择、查询多连接次序、操作站点的选择、数据的传输及数据局部处理等因素,准确地表示了分布式数据库查询执行代价,并且给出了解决基于此代价模型查询优化问题的遗传算法,实验结果表明,该算法有较好的寻优效果,更具有实际意义.A new cost model of distributed database query based on the study of availability semi-joins is put forward in this paper, which can describe the cost of distributed database query precisely by including selection of relation copies, query order of multi-join, determination of performance sites, transaction of data and cost of local computation, and the cost model is analyzed, designed and realized by the constructive global data dictionary and dynamic state dictionary. The genetic algorithm that is used to resolve the query optimization problem based on the new cost model is given, the experiment results indicate that the algorithm is effective.

关 键 词：分布式数据库 查询优化 代价模型 遗传算法 

分 类 号：TP311.133.1[自动化与计算机技术—计算机软件与理论]