东北大学学报（自然科学版）

东北大学学报（自然科学版） ›› 2022, Vol. 43 ›› Issue (5): 609-617.DOI: 10.12068/j.issn.1005-3026.2022.05.001

• 信息与控制 •    下一篇

基于多目标优化的虚拟机部署策略

刘军1， 代福成1， 辛宁2   

  1. (1. 东北大学 计算机科学与工程学院， 辽宁 沈阳110169; 2. 中国空间技术研究院 通导部， 北京100000)
  • 修回日期:2021-01-11 接受日期:2021-01-11 发布日期:2022-06-20
  • 通讯作者: 刘军
  • 作者简介:刘军(1969－)，男，辽宁沈阳人，东北大学教授，博士生导师.
  • 基金资助:
    国家自然科学基金资助项目(61671141).

Virtual Machine Placement Strategy Based on Multi-objective Optimization

LIU Jun1， DAI Fu-cheng1， XIN Ning2   

  1. 1. School of Computer Science & Engineering， Northeastern University， Shenyang 110169， China; 2. Communication Department， China Academy of Space Research， Beijing 100000， China.
  • Revised:2021-01-11 Accepted:2021-01-11 Published:2022-06-20
  • Contact: DAI Fu-cheng
  • About author:-
  • Supported by:
    -

RichHTML

31

PDF (PC)

443

摘要： 为解决虚拟机部署过程中对虚拟机性能、资源利用率、负载均衡值等多个目标的优化问题，提出一种基于强化学习的改进部署算法.首先，用多个目标组成的多维奖励代替原来的单一奖励；然后将资源状态、优化目标及目标占比输入所提的预测器中来预测每个部署方案对应的多维奖励值，并通过反馈结果调节不同优化目标的占比以达到动态多目标优化的目的；最后，为了减少部署时间，用改进的均值聚类算法对服务器资源进行聚类加快部署.通过CloudsimPy平台对算法进行验证，结果表明本文算法可以在相同资源下完成更多的虚拟机请求且具有较高的部署成功率和较低的时延消耗.

关键词: 虚拟机部署;深度强化学习;资源利用率;负载均衡值;虚拟机性能;K均值聚类

Abstract: To optimize the performance， utilization and balance of physical servers in the process of virtual machine placement， an improved placement algorithm based on reinforcement learning is proposed. Firstly， the single reward in reinforcement learning algorithm is replaced by multi-dimensional rewards composed of dynamic objectives， and then the multi-dimensional reward values corresponding to each placement scheme are output by the predictor. The input of the predictor is the resource pool status， optimization target， and target proportion. Finally， the proportions of different optimization objectives are adjusted according to the resource pool status to optimize the targets. The results of CloudsimPy show that the improved algorithm can deploy more virtual machine requests with the higher placement success rate and lower delay consumption under the same resource.

Key words: virtual machine placement; deep reinforcement learning; resource utilization; load balance; virtual machine performance; K-mean clustering

中图分类号: