一种改进的分布式查询优化算法  

Evolutionary algorithm for distributed query optimization

作　　者：于洪涛 钱磊[1] 

机构地区：[1]燕山大学信息科学与工程学院计算机系,河北秦皇岛066004

出　　处：《计算机工程与应用》2013年第8期151-155,共5页Computer Engineering and Applications

摘　　要：为了提高分布式查询优化算法的性能,在遗传模拟退火混合算法中融入小生境技术,并对混合算法的相应要素进行改进,基于该混合算法,提出了一种改进的分布式查询优化算法。利用小生境技术扩展遗传模拟退火混合算法的探索区域,防止早熟现象发生,简化算法中的Meteopolis规则,以消除混合算法中引入新技术后产生的功能冗余,将混合算法应用到分布式查询优化算法中。实验结果表明,改进的分布式查询优化算法可以稳定地得到最优解,减少分布式数据库查询的代价,提高查询效率。In order to improve the performance of distributed query optimization algorithm, the niche technology is introduced into the genetic simulated annealing hybrid algorithm, and some elements of algorithm are improved. An evolutionary algorithm for distributed query optimization is proposed based on the hybrid algorithm. To prevent premature phenomenon, it makes use of niche technology to extend exploration area of genetic simulated annealing algorithm. It simplifies the Meteopolis rule to reduce the redundancy of the new algorithm, the hybrid algorithm is applied to distributed query optimization algorithm. Experimental results show that the evolutionary algorithm for distributed query optimization can obtain excellent result steadily, reduce query cost and improve query efficiency.

关 键 词：小生境遗传模拟退火算法 遗传模拟退火算法 小生境技术 分布式查询 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]