基于改进粒子群算法的云计算平台资源调度  

Resource scheduling of cloud computing platform based on improved particle swarm optimization

作　　者：王德文[1] 刘晓萌[1] 

机构地区：[1]华北电力大学控制与计算机工程学院,河北保定071003

出　　处：《计算机应用研究》2015年第11期3230-3234,3246,共6页Application Research of Computers

基　　金：国家自然科学基金资助项目(61074078);中央高校基本科研业务费专项资金资助项目(12MS113)

摘　　要：针对云计算任务动态变化造成集群资源利用不均衡问题,提出一种基于虚拟机动态迁移技术的云计算资源调度策略。迁移过程中采用引入窗口思想的指数平滑预测确定负载热点,虚拟机选择中综合考虑迁移效果和迁移速度,采用基于退火思想的粒子群算法寻找虚拟机最优放置,并借用轮盘赌思想实现平台资源长期优化。利用云仿真框架Cloud Sim对云计算平台中服务等级协议(SLA)违背率、虚拟机迁移次数、集群能耗以及剩余资源率进行实验,并将本算法与顺序放置、贪心算法和标准粒子群算法进行对比分析,结果表明本算法在上述各方面表现优于其他算法,为提高云计算平台性能提供了新思路。For the unbalanced resource utilization of cloud computing cluster, this paper gave a resource scheduling strategy of cloud based on dynamic migration of virtual machine technology. During the migration process, it firstly determined the hotspots based on index smooth forecasting with window thinking, then selected the virtual machine by considering the effect of＇ migration and migration speed. It used the particle swarm optimization annealing thinking and brag-term optimization goals in the process of migration to search optimal position. By CloudSim simulation framework, the experiment simulated the appearances of the SLA violation rate, the rate of surplus resources, energy and migration times, and the algorithm is better than the greedy algorithm with migration and the standard particle swarm optimization algorithm and the sequence virtual machines placement with non-migratory. Experimental results also show that the algorithm is superior in all respects than the other algorithms, and this algorithm provides a new method on cloud computing platform for the research about improving the performanee of cloud platform.

关 键 词：云计算 虚拟机 动态迁移 模拟退火 粒子群算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]