摘要
基于启发式算法的工作流调度算法目标单一,无法保证用户满意度,且多目标调度算法少、性能差。为了改善现状,提出基于多阶段PSO的多目标工作流调度算法MSPSO,分析工作流任务的层次结构,按层次进行多阶段PSO调度,结合排队理论估算每阶段调度需要的虚拟机数量,控制PSO搜索空间,使算法能快速找到最优解。用4种真实科学工作流在CloudSim环境下进行仿真实验。结果表明,MSPSO算法资源利用率提高了1.81%,能耗降低了9.16%,任务违约率低至0.075%。MSPSO调度算法不仅能动态增减虚拟机,降低能耗,还能在保证截止时间的前提下降低任务违约率,提高资源利用率。
In order to improve the current situation of workflow scheduling algorithm based on heuristic algorithm, which has single goal, can not guarantee user satisfaction, and has few multi-objective scheduling algorithms and poor performance, based on PSO algorithm, a multi-objective workflow scheduling algorithm MSPSO based on multi-stage PSO is proposed. Firstly, the hierarchical structure of workflow tasks is analyzed, and multi-stage PSO scheduling is carried out according to the hierarchical structure. At the same time, the number of virtual machines needed for each stage scheduling is estimated by queuing theory, and the search space of PSO is controlled, so that the algorithm can quickly find the optimal solution. Four kinds of real scientific workflow are used to simulate the experiment in CloudSim environment. The results show that the resource utilization of MSPSO algorithm is increased by 1.81%, the energy consumption is reduced by 9.16%, and the task default rate is reduced to 0.075%. MSPSO scheduling algorithm can not only dynamically increase and decrease the energy consumption of virtual machines, but also reduce the default rate of tasks and improve resource utilization on the premise of guaranteeing deadlines.
作者
申秋慧
牛玲
SHEN Qiu-hui;NIU Ling(School of Computer Science and Technology,Zhoukou Normal University,Zhoukou 466000,China)
出处
《软件导刊》
2019年第9期81-84,共4页
Software Guide
基金
河南省高校科技创新团队项目(17IRTSTHN009)
河南省教育厅重点研究项目(16A520104)
