基于布谷鸟搜索优化马尔可夫的文件热度预测  

Optimizing file temperature prediction of Markov based on cuckoo search

作　　者：王彪 王克俭[1] 何振学 高万豪 魏雪川 WANG Biao;WANG Ke-jian;HE Zhen-xue;GAO Wan-hao;WEI Xue-chuan(College of Information Science and Technology,Hebei Agricultural University,Baoding 071001,China;College of Control and Mechanical Engineering,Tianjin Chengjian University,Tianjin 300384,China;Shijiazhuang Power Supply Company,State Grid Hebei Electric Power Corporation,Shijiazhuang 050052,China)

机构地区：[1]河北农业大学信息科学与技术学院,河北保定071001 [2]天津城建大学控制与机械工程学院,天津300384 [3]国网河北省电力公司石家庄供电公司,河北石家庄050052

出　　处：《计算机工程与设计》2021年第11期3121-3127,共7页Computer Engineering and Design

基　　金：河北省教育厅重点研究基金项目(ZD2016158);河北省高可信信息系统重点实验室开放研究课题基金项目(2018KFKT003);河北省高等学校科学技术研究基金项目(BJ2019008);河北省自然科学基金项目(F2020204003)。

摘　　要：针对现有分布式文件系统中静态副本管理机制的不足,提出一种基于布谷鸟搜索优化马尔可夫模型的算法进行文件热度预测。根据文件最近的访问特征采用无偏灰色模型预测下一时刻文件访问热度,利用布谷鸟搜索优化马尔可夫模型,对预测结果进行误差修正,根据新陈代谢思想对预测序列进行等维处理,通过最新数据趋势得到预测结果。实验结果表明,MAPE值达到3.08%,与其它模型相比,MAPE值平均降低了2.26%。该方法对文件热度预测达到了较好结果,为文件副本的动态管理提供依据。Aiming at the shortcomings of the static copy management mechanism in the existing distributed file system,an algorithm based on the cuckoo search optimization Markov model was proposed to predict file popularity.The unbiased gray model was used to predict the file access heat at the next moment according to the recent access characteristics of the file,the cuckoo search was used to optimize the Markov model,and the error of the prediction result was corrected.The prediction sequence was equal-dimensional according to the metabolism idea process,the forecast result was got through the latest data trend.Experimental results show that the MAPE value reaches 3.08%,which is an average reduction of 2.26%compared with other models.It is verified that this method achieves good results in prediction of document popularity and provides a basis for the dynamic management of document copies.

关 键 词：热度预测 无偏灰色预测 马尔可夫模型 新陈代谢 布谷鸟搜索 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]