摘要
针对多QoS约束的工作流调度优化问题,提出一种云环境下的工作流多目标调度遗传算法。为寻找满足预算和期限双QoS约束的工作流执行时间与代价的同步最优解,建立遗传调度模型。对工作流进行结构分层,确保任务执行顺序;将个体惩罚因素引入适应度函数中,保留接近约束的边界解,扩展解空间分布;在层次、任务及资源层面上设计3种遗传交叉和变异,增加最优解的求解概率。实验结果表明,该算法在Pareto最优解分布、收敛性和稳定性及调度效率方面均优于基准算法。
Aiming at the optimization problem of workflow scheduling under multi-QoS constraints,a cloud workflow multi-objective scheduling genetic algorithm was designed.For searching the synchromization optimization solutions of workflow execution time and cost satisfying bi-QoS of budget and deadline,the genetic scheduling model was built.The workflow structure was divided into several levels,which ensured the execution order.The individual punishment element was introduced into the fitness function,which kept the boundary solution close to constraint and extended the solution space.Three kinds of genetic crossover and mutation operation were respectively introduced among levels,tasks and resources,which improved the solving rate of the optimal solution.Experimental results show that the proposed algorithm performs better than other baseline algorithms in the distribution of Pareto optimal,convergence and stability and scheduling efficiency.
作者
薛庆水
李凤英
XUE Qing-shui;LI Feng-ying(School of Computer Science and Information Engineering,Shanghai Institute of Technology,Shanghai 201418,China;School of Continuing Education,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《计算机工程与设计》
北大核心
2019年第8期2196-2203,共8页
Computer Engineering and Design
基金
国家自然科学基金项目(61170227)
上海应用技术大学协同创新基金项目(39120K178038)
关键词
云计算
工作流调度
遗传算法
任务分层
多目标优化
cloud computing
workflow scheduling
genetic algorithm
tasks leveling
multi-objective optimization