基于遗传算法的批处理科学工作流任务调度算法的改进

Improving Task Scheduling of Batch Workflow Applications Based on Genetic Algorithm

随着云计算应用的不断深入以及对大数据处理需求的不断提升,越来越多的企业选择使用云平台处理海量的数据。由于云计算的商业性,这就对云计算中的任务调度提出了更加严苛的要求,如何合理且经济地完成任务调度成为了研究云计算的关键问题之一。批处理科学工作流是大数据时代的一种新型工作流建模形式,近两年已引起业内的重视,但当前仍处于起步阶段。本文首先对当前传统的任务调度算法进行分析,并指出其中的不足之处,从而改进了基于遗传算法的批处理科学工作流任务调度算法BIGA(batch scientific workflow task scheduling based on improved genetical gorithms),在满足固定截止期的条件下,以任务调度成本最优优化目标,分别对独立任务调度与非独立任务调度进行研究实验。最后在Matlab中进行模拟实验,结果表明:本文的改进算法在满足任务截止期的情况下与按比例划分截止期经典调度算法相比,在一定任务规模下,完成任务调度所需成本更低,更加符合云资源的使用特征与用户需求。

With the continuous development of cloud computing application and the increasing demand for big data processing,more and more enterprises choose to use the cloud platform to process massive amounts of data.The commercial nature of cloud computing is giving task scheduling in cloud computing more stringent requirements,and how to reasonably and economically complete the task scheduling is one of the key problems of cloud computing.Batch scientific workflow is a new type of workflow model in big data era,which has attracted the attention of industry in the past two years,but is still in its infancy.This article analyzes the current task scheduling algorithm,points out its shortcomings,and then puts forward a new batch BIGA task scheduling algorithm based on genetic algorithm(batch scientific workflow task scheduling based on improved genetic algorithms).Under the condition of meeting the deadline and aimed at task scheduling cost optimization,experiments of both independent task scheduling and dependent task scheduling were conducted.Finally,in Matlab simulation experiments,the results show that the improved BIGA algorithm,compared with the classic task scheduling algorithm,has the advantage of lower cost and is more in line with the cloud resources characteristics.It can also better meet the user’ requirements.