当前位置: X-MOL 学术Neural Comput. & Applic. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
A hybrid genetic algorithm for scientific workflow scheduling in cloud environment
Neural Computing and Applications ( IF 6 ) Pub Date : 2020-05-11 , DOI: 10.1007/s00521-020-04878-8
Hatem Aziza , Saoussen Krichen

Nowadays, we live an unprecedented evolution in cloud computing technology that coincides with the development of the vast amount of complex interdependent data which make up the scientific workflows. All these circumstances developments have made the issue of workflow scheduling very important and of absolute priority to all overlapping parties as the provider and customer. For that, work must be focused on finding the best strategy for allocating workflow tasks to available computing resources. In this paper, we consider the scientific workflow scheduling in cloud computing. The main role of our model is to optimize the time needed to run a set of interdependent tasks in cloud and in turn reduces the computational cost while meeting deadline and budget. To this end, we offer a hybrid approach based on genetic algorithm for modelling and optimizing a workflow-scheduling problem in cloud computing. The heterogeneous earliest finish time (HEFT), an heuristic model, intervenes in the generation of the initial population. Based on results obtained from our simulations using real-world scientific workflow datasets, we demonstrate that the proposed approach outperforms existing HEFT and other strategies examined in this paper. In other words, experiments show high efficiency of our proposed approach, which makes it potentially applicable for cloud workflow scheduling. For this, we develop a GA-based module that was integrated to the WorkflowSim framework based on CloudSim.



中文翻译:

云环境下科学工作流调度的混合遗传算法

如今,我们在云计算技术方面经历了前所未有的发展,这与构成科学工作流程的大量复杂的相互依存数据的发展相吻合。所有这些情况的发展使得工作流调度的问题变得非常重要,并且对于作为提供者和客户的所有重叠方都具有绝对优先权。为此,工作必须集中于寻找最佳的策略,以将工作流任务分配给可用的计算资源。在本文中,我们考虑了云计算中科学的工作流调度。我们的模型的主要作用是优化在云中运行一组相互依赖的任务所需的时间,进而减少计算成本,同时满足截止日期和预算。为此,我们提供了一种基于遗传算法的混合方法,用于对云计算中的工作流计划问题进行建模和优化。异类最早完成时间(HEFT)是一种启发式模型,会干预初始人口的产生。基于使用现实世界中科学工作流数据集进行的仿真得出的结果,我们证明了所提出的方法优于现有的HEFT和本文研究的其他策略。换句话说,实验表明我们提出的方法具有很高的效率,这使其有可能适用于云工作流调度。为此,我们开发了一个基于GA的模块,该模块已集成到基于CloudSim的WorkflowSim框架中。干预初始人口的产生。基于使用现实世界中科学工作流数据集进行的仿真得出的结果,我们证明了所提出的方法优于现有的HEFT和本文研究的其他策略。换句话说,实验表明我们提出的方法具有很高的效率,这使其有可能适用于云工作流调度。为此,我们开发了一个基于GA的模块,该模块已集成到基于CloudSim的WorkflowSim框架中。干预初始人口的产生。基于使用现实世界中科学工作流数据集进行的仿真得出的结果,我们证明了所提出的方法优于现有的HEFT和本文研究的其他策略。换句话说,实验表明我们提出的方法具有很高的效率,这使其有可能适用于云工作流调度。为此,我们开发了一个基于GA的模块,该模块已集成到基于CloudSim的WorkflowSim框架中。这使其有可能适用于云工作流调度。为此,我们开发了一个基于GA的模块,该模块已集成到基于CloudSim的WorkflowSim框架中。这使其有可能适用于云工作流调度。为此,我们开发了一个基于GA的模块,该模块已集成到基于CloudSim的WorkflowSim框架中。

更新日期:2020-05-11
down
wechat
bug