基于改进型PSO的成本最小化云任务调度方案  

Cloud task scheduling algorithm to minimize cost based on improved PSO

作　　者：黄彩娟[1] 刘卓华[1] 郑荣茂[1] 杨滨[2] 

机构地区：[1]广东机电职业技术学院信息工程学院,广东广州510515 [2]江南大学设计学院,江苏无锡214122

出　　处：《计算机工程与设计》2017年第12期3349-3353,3401,共6页Computer Engineering and Design

基　　金：广东省高等学校优秀青年教师培养计划基金项目(Yq2013171);强师工程职业教育教师技能提升工程基金项目(YJKJGZS2015-04)

摘　　要：针对云计算环境中工作流的成本有效调度问题,提出一种基于改进型粒子群优化(PSO)算法的调度方法。根据虚拟机(VM)资源的单位成本,对其进行排序和编号,提高PSO的学习效率。在满足任务截止时间约束下,以最小化成本为目标,执行PSO算法来获得调度解;通过任务的数据传输时间矩阵和执行时间矩阵计算VM的租用时间,精确地给出解的适应度值,提高PSO的求解精度。仿真结果表明,该算法收敛速度快,获得的调度方案具有较低的执行成本。For the cost effective scheduling issue of workflow in the cloud computing environment,a scheduling method based on improved particle swarm optimization(PSO)algorithm was proposed.The virtual machine(VM)resources were renumbered according to the unit cost,so as to improve the learning efficiency of PSO.The PSO algorithm was used to obtain the scheduling solution which minimized the cost and met the deadline of the task.The data transfer time matrix and execution time matrix were built to calculate the leased time of VM,and the fitness value of solution was then obtained,so as to improve the accuracy of the PSO solution.Simulation results show that the proposed algorithm has fast convergence speed,and the scheduling scheme has lower execution cost.

关 键 词：云环境 任务调度 最小化成本 虚拟机编号 粒子群优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]