云计算环境下基于微粒群的虚拟机任务调度算法  

Virtual Machine Task Scheduling Algorithm Based on PSO in Cloud Computing Environment

作　　者：胡德敏[1,2] 户静[1] 余星[1] 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]上海理工大学软件教研室,上海200093

出　　处：《计算机测量与控制》2014年第4期1189-1192,共4页Computer Measurement &Control

基　　金：国家自然科学基金项目(61202376);上海市教委科研创新项目(13YZ075)

摘　　要：为提高云计算环境中虚拟机任务调度的执行效率和充分发挥云计算技术优势，提出一种基于微粒群的虚拟机任务调度算法——PSOTS算法；PSOTS算法以完成任务最短时间为目标，首先通过设计一种新型的编码方式使得连续的微粒群算法适用于离散的虚拟机任务调度问题；然后引入禁忌搜索算法增强种群的多样性以避免微粒因早熟而陷于局部最优的问题；实验证明，在50～500个任务的情况下，PSOTS算法调度效率优于Min—min算法和遗传算法（GA），有效缩短任务执行时间和解决云环境下虚拟机任务调度问题。This paper proposed a virtual machine tasks scheduling algorithm named PSOTS that based on particle swarm algorithm to im- prove tasks scheduling execution efficiency of virtual machines in the cloud computing environment and to give full play to the advantage o~ cloud computing technology. PSOTS algorithm takes shortest time of task execution as goal. First, try to make the continuous particle swarm algorithm applied to discrete virtual machine tasks scheduling problem by a new encoding method; Then the TS algorithm is intro- duced to enhance the diversity of group and avoid particles being trapped in local optimum caused by premature convergence; Finally, ac- cording to the experiment, In the case of 50--500 tasks, it turns out that the efficiency of PSOTS scheduling is superior to the Min--min al- gorithm and GA algorithm. PSOTS algorithm effectively reduces the execution time of tasks and solves the problem of VM tasks scheduling in the cloud computing environment.

关 键 词：云计算 虚拟机 微粒群算法 禁忌搜索算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]