计算机集成制造系统 ›› 2013, Vol. 19 ›› Issue (08 ): 1849-1858.

• 产品创新开发技术 • 上一篇    下一篇

面向价格动态变化的云工作流调度算法

郑敏,曹健,姚艳   

  1. 上海交通大学计算机科学与工程系协同计算与智能技术实验室
  • 出版日期:2013-08-31 发布日期:2013-08-31
  • 基金资助:
    国家自然科学基金资助项目(61073021,61272438);上海市科委资助项目(12511502704,11511500102,10DZ1200200);上海交通大学医工交叉基金资助项目(YG2011MS38)。

Cloud workflow scheduling algorithm oriented to dynamic price changes

  • Online:2013-08-31 Published:2013-08-31
  • Supported by:
    Project supported by the National Natural Science Foundation,China (No.61073021,61272438),the Science and Technology Commission of Shanghai Municipality,China (No.12511502704,11511500102,10DZ1200200) and the Cross Research Fund of Biomedical Engineering of Shanghai Jiaotong University,China (No.YG2011MS38).

摘要: 为解决资源价格变动环境中云工作流任务调度开销优化的问题,提出一种基于动态规划的云工作流资源调度算法。该算法将云工作流中的任务分为多个串行任务组,然后利用动态规划算法对每组任务进行调度,计算在该任务组规定完成时间内开销最低的调度方案。通过实验,与状态图搜索算法进行了比较,在被调度的资源的价格为可变函数时,计算了四种工作流模型的调度结果。结果表明,所提动态规划算法比考虑资源平均价格的状态图搜索算法有更好的表现。

关键词: 工作流, 云计算, 调度, 动态规划, 资源价格

Abstract: To solve the scheduling overhead optimization of cloud workflow in dynamic resources prices environment,a cloud workflow algorithm based on dynamic planning was introduced.The cloud workflow task was divided into groups of serial work by this algorithm and the dynamic algorithm was used to dispatch each task.The lowest overhead scheduling scheme in certain time was calculated.The proposed algorithm was compared with state space algorithm by experiments,and four kinds of workflow models'scheduling results was computed when the resource price as variable function.The result showed that the dynamic planning algorithm had better performance than state space algorithm with considering the changing price.

Key words: workflow, cloud computing, scheduling, dynamic planning, resource price

中图分类号: