摘要： 基于数据垂直划分的分布并行Skyline查询算法大多并行性较低，无法适应海量分布式数据的快速响应要求。为此，在BDS算法的基础上提出一种更高效的分布并行Skyline查询算法PDS-VP。其中，节点被分为协调者与参与者，原本由协调者节点完成的随机访问和本地Skyline计算分发给各参与者节点进行处理，以提高算法的执行效率。实验结果证明，该算法提高了原算法的并行性和运行效率。

关键词: Skyline查询, 分布式环境, 并行Skyline, 数据垂直划分, 多目标优化, 数据挖掘

Abstract: Most distributed and parallel Skyline query algorithms based on data vertical partition have poor parallelism, which makes them inadaptable to the queries on massive data with fast response requirement. This paper proposes an effective distributed and parallel Skyline query algorithm named PDS-VP(Parallel and Distributed Skyline query for Vertical Partitioning datasets). There are two kinds of nodes in PDS-VP: the coordinator and the participant. Tasks of the random access and Skyline computation in locals in the coordinator are assigned to the participants to enhance the parallelism, so as to improve the efficiency of the algorithms. Experimental results show that PDS-VP has higher parallelism and is more effective than the existing methods.

Key words: Skyline query, distributed environment, parallel Skyline, data vertical partition, multi-object optimization, data mining

中图分类号: 

• TP391