摘要
针对海量输变电工程数据检索效率低下的问题,提出面向输变电工程元数据的分级索引算法,结合灰度预测模型对元数据热度进行预测,将元数据划分为热点与非热点两级。在此基础上,对热点元数据依据预测热度进行分区并分别建立B+树索引,对非热点元数据以相同方式分区并分别选取相应哈希函数,在保证较高检索效率的同时,减小空间开销。通过实验数据集进行测试与分析,其结果表明,该方法能够有效减小查询时的时空开销,实现对输变电工程数据的高效检索。
Aiming at the problem of low data retrieval efficiency of massive power transmission and transformation engineering,a hierarchical indexing algorithm for metadata of power transmission and transformation engineering was proposed.Combined with gray prediction model to predict the heat of metadata,the metadata was divided into two levels,hotspots and non-hotspots.On this basis,the hotspot metadata was partitioned according to the predicted heat and the B+tree index was established respectively.The non-hotspot metadata was partitioned in the same way and the corresponding hash function was selected respectively to ensure the high retrieval efficiency and reduce the space.Overhead test and analysis were carried out through experimental data sets.Experimental results show that the proposed method can effectively reduce the space-time overhead during query and realize efficient retrieval of data in power transmission and transformation engineering.
作者
孙小虎
宋慧娟
代安琪
许刚
SUN Xiao-hu;SONG Hui-juan;DAI An-qi;XU Gang(Data Center,State Grid Economic and Technological Research Institute Limited Company,Beijing 102209,China;Operation and Maintenance Department,State Grid Shanghai Municipal Electric Power Company,Shanghai 200122,China;School of Electrical and Electronic Engineering,North China Electric Power University,Beijing 102206,China)
出处
《计算机工程与设计》
北大核心
2019年第11期3192-3199,共8页
Computer Engineering and Design
基金
国家电网有限公司科技基金项目"输变电工程数字化设计成果在运维管理中的研究应用"(B3441617K003)
关键词
输变电工程
分级管理
元数据索引
海量数据
灰度预测
power transmission and transformation project
hierarchical management
metadata index
massive data
grey prediction