改进CS算法结合决策树的云工作流调度

发布时间：2018-10-12 17:50

【摘要】：对云计算环境下工作流任务调度的现有方案进行分析,针对存在运行时间长、资源利用率低等不足,提出一种结合改进型布谷鸟搜索算法和决策树的工作流任务调度方案。首先,根据工作流任务属性分配截止期限;其次,利用改进型布谷鸟搜索算法将工作流分割成多个子工作流,最小化数据依赖性,再利用决策树选择出满足任务Qo S约束的资源;最后,根据任务的计算时间、排队时间和通信延迟的总和来判断是否满足截止期限约束,以此配置相应的资源。实验结果表明,该方案具有较短的总运行时间和较高的任务完成率。
[Abstract]:This paper analyzes the existing scheme of workflow task scheduling in cloud computing environment and proposes a workflow task scheduling scheme combining improved cuckoo search algorithm and decision tree in view of the shortcomings such as long running time and low resource utilization. First, the deadline is assigned according to the task attributes of workflow. Secondly, the improved cuckoo search algorithm is used to divide the workflow into several sub-workflows to minimize the data dependency. Then the decision tree is used to select the resources that meet the Qo S constraints of the task. Finally, according to the computation time of the task, the sum of the queuing time and the communication delay, it is judged whether the deadline constraints are met or not, and the corresponding resources are configured accordingly. The experimental results show that the scheme has shorter total running time and higher task completion rate.
【作者单位】：四川理工学院计算机学院;
【基金】：四川省教育厅重点项目(15ZA0224) 人工智能四川省重点实验室(2014RYJ01) 四川省智慧旅游研究基地规划项目(ZHZ14-01)
【分类号】：TP18;TP3

【相似文献】