东北大学学报:自然科学版 ›› 2017, Vol. 38 ›› Issue (2): 209-213.DOI: 10.12068/j.issn.1005-3026.2017.02.012

• 信息与控制 • 上一篇    下一篇

基于深度置信网络的云应用负载预测方法

马安香, 张长胜, 张斌, 张晓红   

  1. (东北大学 计算机科学与工程学院, 辽宁 沈阳110169)
  • 收稿日期:2015-09-16 修回日期:2015-09-16 出版日期:2017-02-15 发布日期:2017-03-03
  • 通讯作者: 马安香
  • 作者简介:马安香(1979-),女,辽宁沈阳人,东北大学讲师,博士; 张斌(1964-),男,辽宁本溪人,东北大学教授,博士生导师.
  • 基金资助:
    国家科技支撑计划项目(2014BAI17B00); 国家自然科学基金资助项目(61572116,61572117,61502089).

Load Prediction Approach for Cloud Application Based on Deep Belief Networks

MA An-xiang, ZHANG Chang-sheng, ZHANG Bin, ZHANG Xiao-hong   

  1. School of Computer Science & Engineering, Northeastern University, Shenyang 110169, China.
  • Received:2015-09-16 Revised:2015-09-16 Online:2017-02-15 Published:2017-03-03
  • Contact: ZHANG Bin
  • About author:-
  • Supported by:
    -

摘要: 为了准确预测云应用负载以便及时执行云应用自适应优化,从而保证云应用性能的稳定,根据云环境下应用负载预测问题的特点,提出了基于深度置信网络的云应用负载预测方法.首先给出能够有效描述负载数据的显式特征和隐式特征并定义了负载预测模型,进而给出基于深度置信网络的负载预测算法.对算法进行了分析并在真实数据集上与相关算法进行了比较,结果表明,本文提出的方法能够更加有效地解决云应用负载预测问题.

关键词: 云计算, 云应用, 深度置信网络, 负载预测, 自适应优化

Abstract: To implement the adaptive optimization to ensure the performance of cloud application, it is necessary to accurately predict the load for cloud application. According to the feature of load prediction in cloud application, an approach is proposed for load prediction based on deep belief networks. Explicit and implicit features for load data are given. Load prediction model is defined. Then, the algorithm of load prediction based on deep belief networks is designed and implemented. This approach is evaluated and compared with some related load prediction algorithms, which reveals very encouraging results in terms of the prediction quality.

Key words: cloud computing, cloud application, deep belief network, load prediction, adaptive optimization

中图分类号: