查询结果:   方军,张璋,张雪峰,杜聪,马涛.基于均衡适应度的云工作流调度算法[J].计算机应用与软件,2019,36(5):255 - 261.
中文标题
基于均衡适应度的云工作流调度算法
发表栏目
算法
摘要点击数
319
英文标题
CLOUD WORKFLOW SCHEDULING ALGORITHM BASED ON TRADE-OFF FITNESS
作 者
方军 张璋 张雪峰 杜聪 马涛 Fang Jun Zhang Zhang Zhang Xuefeng3 Du Cong Ma Tao
作者单位
国家质量监督检验检疫总局信息中心 北京 100088 北京中质信维科技有限公司 北京 100088 北京赛迪工业和信息化工程监理中心有限公司 北京 100048 石家庄铁道大学信息学院 河北 石家庄 050043  
英文单位
Information Center, General Administration of Quality Supervision, Inspection and Quarantine, Beijing 100088, China Beijing Zhong Xin Wei Technology Co., Ltd., Beijing 100088, China Beijing Sai Di Industrial and Information Engineering Supervision Center Co., Ltd., Beijing 100048, China School of Information,Shijiazhuang Tiedao University, Shijiazhuang 050043, Hebei, China  
关键词
云计算 任务调度 遗传算法 执行代价 调度效率
Keywords
Cloud computing Task scheduling Genetic algorithm Execution cost Scheduling efficiency
基金项目
国家自然科学基金项目(11702179)
作者资料
方军,高工,主研领域:云平台信息化管理。张璋,工程师。张雪峰,高工。杜聪,工程师。马涛,高工。刘鑫,讲师。 。
文章摘要
针对工作流任务调度优化问题,提出一种云工作流任务调度遗传算法。为了寻找工作流执行时间与执行代价的同步最优解,建立了遗传调度模型。在个体编码方面,采用了一种二维排列编码方法,可以更好地展现工作流任务间的执行次序;综合考虑任务执行代价与最早完成时间两个因素,设计了一种均衡适应度函数;为了丰富种群个体多样性,引入三种遗传交叉操作和两种遗传变异操作,以产生新的个体,增加了最优解的求解概率。通过数值仿真实验,在多个性能指标上对算法进行分析。结果表明,该调度算法能更好地平衡执行代价与调度效率,性能优于同类算法。
Abstract
For the optimization of workflow task scheduling, we presented the cloud workflow task scheduling genetic algorithm. To search the synchronous optimal solution of workflow execution time and cost, we built the genetic scheduling model. In the individual coding, two-dimension array coding method was adopted, which could better show the execution order between workflow tasks. Considering both the task execution cost and earliest finish time, we designed the trade-off fitness function. For enriching the diversity of the population individuals, we introduced three kinds of genetic crossover operation and two kinds of genetic mutation operation, which generated new individuals and increased the probability of obtaining the optimal solution. Through the numerical simulation experiments, we analyzed the algorithm in multiple performance indexes. The results show that the scheduling algorithm can achieve a better balance between the execution cost and the scheduling efficiency, and its performance is superior to that of similar algorithms.
下载PDF全文