集群环境下矢量空间数据长期存储方法仿真
|
摘要
针对当前数据存储方法存在存储空间占用严重和实时性差的问题,提出基于粒子群算法和K均值的集群环境下矢量空间数据长期存储方法。依据矢量空间数据比例尺和格网层次之间的关系梳理,利用Hilbert填充曲线划分矢量空间数据。在得出划分结果的基础上,将数据要素Hilbert排列码、矢量空间数据要素多级格网编码、数据要素顺序码、数据要素标识码,完成矢量空间数据唯一标识操作。确定K均值聚类算法初始均值或中心位置之后,初始化粒子群算法运行参数,计算粒子群适应度值,并更新粒子移动速度与位置,以此得到最佳粒子位置,将其当作K均值聚类的最佳聚类中心。将具有唯一标识的矢量空间数据集代入聚类算法中,并将具有相似性的数据缓存至同一数据簇,实现矢量空间数据长期存储。实验结果表明,所提方法数据存储空间占用率较低,存储效率高,具有可行性。
This article presented a long-term storage method of vector space data in clustered environment based on particle swarm optimization and K-means. According to the relationship between the vector space data scale and the grid level, Hilbert filling curve was used to divide the data on vector space. Based on the division result, the data element Hilbert permutation code, the multi-level grid coding of vector spatial data element, the data element sequence code and the data element identification code were used to complete the operation for unique identification of vector spatial data. After determining the initial mean value or central position of K-means clustering algorithm, the operating parameters of particle swarm algorithm were initialized, the particle swarm fitness value was calculated, and then the particle moving speed and position of particle were updated, and thus the best particle position was obtained, which could be taken as the best clustering center K-means cluster. Finally, the data set of vector space with unique identifier was introduced into the clustering algorithm. Meanwhile, the data with similarity was cached to the same data cluster. Thus, the long-term storage of data in vector space was achieved. Simulation results verify that the proposed method has low data storage space and high storage efficiency, which is feasible.
This article presented a long-term storage method of vector space data in clustered environment based on particle swarm optimization and K-means. According to the relationship between the vector space data scale and the grid level, Hilbert filling curve was used to divide the data on vector space. Based on the division result, the data element Hilbert permutation code, the multi-level grid coding of vector spatial data element, the data element sequence code and the data element identification code were used to complete the operation for unique identification of vector spatial data. After determining the initial mean value or central position of K-means clustering algorithm, the operating parameters of particle swarm algorithm were initialized, the particle swarm fitness value was calculated, and then the particle moving speed and position of particle were updated, and thus the best particle position was obtained, which could be taken as the best clustering center K-means cluster. Finally, the data set of vector space with unique identifier was introduced into the clustering algorithm. Meanwhile, the data with similarity was cached to the same data cluster. Thus, the long-term storage of data in vector space was achieved. Simulation results verify that the proposed method has low data storage space and high storage efficiency, which is feasible.
引文
[1] 王彦佐,周伟,冯磊.内存数据库在ZY1-02C海量数据空间检索中的应用[J].国土资源遥感,2018,30(1):238-242.
[2] 毛健,朱长青,张兴国,等.一种矢量地理空间数据文件细粒度访问控制模型[J].地理与地理信息科学,2017,33(1):13-18.
[3] 褚龙现,李晓英,陈旭,等.一种支持高效并行处理的矢量数据索引方法[J].计算机工程与应用,2017,53(11):79-84.
[4] 孙群.多源矢量空间数据融合处理技术研究进展[J].测绘学报,2017,46(10):1627-1636.
[5] 徐英辉,祝恩国,赵睿,等.MongoDB索引的用电信息非结构化数据存储方法[J].电力系统及其自动化学报,2017,29(9):93-97.
[6] 陈正旭,李爽爽,孙晓燕.一种基于NoSQL的气象非结构化数据产品存储方法[J].气象科技,2017,45(3):430-434.
[7] 陈重韬,王伟平,孟丹,等.面向海量NetFlow数据的存储和查询处理方法研究[J].高技术通讯,2016,26(6):534-541.
[8] 胡皓然,王颖.结合HSV空间阴影消除的高斯混合模型背景减除方法[J].电子设计工程,2018,26(7):31-34.
[9] 何诚刚.大规模电子通信信息存储效率管理仿真[J].计算机仿真,2017,34(9):175-178.
[10] 马子泰,曹健,姚艳.云环境下使用竞价实例并考虑中间数据存储策略的工作流调度方法[J].计算机集成制造系统,2017,23(5):983-992.
[11] 朱永利,李莉,宋亚奇,等.ODPS平台下的电力设备监测大数据存储与并行处理方法[J].电工技术学报,2017,32(9):199-210.
[2] 毛健,朱长青,张兴国,等.一种矢量地理空间数据文件细粒度访问控制模型[J].地理与地理信息科学,2017,33(1):13-18.
[3] 褚龙现,李晓英,陈旭,等.一种支持高效并行处理的矢量数据索引方法[J].计算机工程与应用,2017,53(11):79-84.
[4] 孙群.多源矢量空间数据融合处理技术研究进展[J].测绘学报,2017,46(10):1627-1636.
[5] 徐英辉,祝恩国,赵睿,等.MongoDB索引的用电信息非结构化数据存储方法[J].电力系统及其自动化学报,2017,29(9):93-97.
[6] 陈正旭,李爽爽,孙晓燕.一种基于NoSQL的气象非结构化数据产品存储方法[J].气象科技,2017,45(3):430-434.
[7] 陈重韬,王伟平,孟丹,等.面向海量NetFlow数据的存储和查询处理方法研究[J].高技术通讯,2016,26(6):534-541.
[8] 胡皓然,王颖.结合HSV空间阴影消除的高斯混合模型背景减除方法[J].电子设计工程,2018,26(7):31-34.
[9] 何诚刚.大规模电子通信信息存储效率管理仿真[J].计算机仿真,2017,34(9):175-178.
[10] 马子泰,曹健,姚艳.云环境下使用竞价实例并考虑中间数据存储策略的工作流调度方法[J].计算机集成制造系统,2017,23(5):983-992.
[11] 朱永利,李莉,宋亚奇,等.ODPS平台下的电力设备监测大数据存储与并行处理方法[J].电工技术学报,2017,32(9):199-210.