多维QoS约束的云计算工作流调度算法
Cloud workflow scheduling algorithm with multidimensional QoS constraints
任小强 1聂清彬 1姜慧 1王浩宇1
作者信息
- 1. 西南交通大学希望学院信息工程系,四川成都 610400
- 折叠
摘要
为有效解决云计算异构系统中工作流调度问题,提出一种多维QoS约束下的改进遗传算法(QoS-IGA).建立工作流任务调度模型、多维QoS约束模型和考虑任务完成时间、完成费用及虚拟资源可靠性和负载均衡度的多目标优化函数;提出种群初始化原则,以及不破坏任务间依赖关系的交叉与变异算子,引入模拟退火算法的Metropolis准则避免遗传算法的早熟收敛问题.实验结果表明,QoS-IGA算法有效可行,其收敛速度快,调度效率高.
Abstract
To effectively solve the workflow scheduling problem in cloud computing heterogeneous systems,an improved genetic algorithm(QoS-IGA)under multidimensional QoS constraints was proposed.A workflow task scheduling model,a multidimen-sional QoS constraint model and a multi-objective optimization function were established considering task completion time,com-pletion cost and virtual resource reliability and load balancing degree,and a population initialization principle was proposed,as well as crossover and variation operators that did not destroy inter-task dependencies,and the Metropolis criterion of simulated annealing algorithm was introduced to avoid the premature convergence problem.Experimental simulation results show that the QoS-IGA algorithm is effective and feasible,and its convergence speed and the scheduling efficiency are both high.
关键词
云计算/服务质量/遗传算法/工作流调度/有向无环图/负载均衡/模拟退火算法Key words
cloud computing/quality of service/genetic algorithms/workflow scheduling/directed acyclic graphs/load balancing/simulated annealing algorithms引用本文复制引用
基金项目
2022年教育部高教司供需对接就业育人基金项目(20230104390)
2022年教育部高教司供需对接就业育人基金项目(20230106769)
成都市交通+旅游大数据应用技术研究基金项目(2021003)
西南交通大学希望学院一流本科课程建设基金项目(2112056)
西南交通大学希望学院一流本科课程建设基金项目(2022107)
出版年
2024
NETL
NSTL
万方数据