摘要： 传统蚁群算法在解决数据仓库查询优化问题时存在过早收敛、收敛速度慢的缺点。为此，对传统蚁群算法进行改进，将伪随机状态转移规则引入最大最小蚁群系统，在每次迭代结束后进行迭代局部搜索。实验结果表明，改进算法在多表连接查询优化中具有较快的收敛速度，能提高最优解的质量。

关键词: 蚁群算法, 迭代局部搜索, 数据仓库, 多连接查询优化, 查询执行计划

Abstract: Traditional Ant Colony Algorithm(ACA) is applied to solve the query optimization problem of Data Warehouse(DW), it has some shortcomings such as premature convergence and slowly convergence. This paper improves the traditional ACA to address these issues. The pseudo-random proportion rule is introduced to the Max-Min Ant System(MMAS), and the Iterated Local Search(ILS) strategy is performed after each iteration. Experimental results show that the improved algorithm accelerates the convergence rate of the algorithm and improves the quality of the optimal solution in solving multi-join query optimization.

Key words: Ant Colony Algorithm(ACA), Iterated Local Search(ILS), Data Warehouse(DW), multi-join query optimization, Query Execution Plan(QEP)

中图分类号: 

• TP391