摘要
虚拟化技术为云环境下的商业化服务提供了技术发展平台.针对虚拟机资源调度问题,以往的研究提出面向用户满意度或面向Qo S的调度算法.前者缺乏物理资源利用率的考虑,后者忽略了用户的体验满意度.为了解决这些不足之处,提出多目标PSO优化算法的云服务策略(U-RPSO算法).该算法从用户满意度和资源利用率多目标问题着手,通过引入聚集变化因子动态改变粒子种群更新机制.为了验证U-RPSO算法,扩展cloudsim包模拟云环境进行仿真实验,并与改进的粒子群算法和双适应度算法进行对比测试,实验结果显示,该算法不仅提高了用户任务的完成效率,同时降低了用户任务的成本耗费,并充分利用了云环境下的物理资源,使得资源利用率最大化.
Virtualization technology provides a platform for business services in the cloud environment. For the problem of virtual ma- chine resource scheduling, previous studies proposed scheduling algorithrn for user satisfaction or QoS. The former lacks the considera- tion of physical resource utilization and the latter neglects the user's experience satisfaction. To overcome these drawbacks, we propose cloud service strategy for multi objective PSO algorithm ( U-RPSO algorithm }. The algorithm designs the virtual machine allocation strategy which is based on the problem of user satisfaction and resource utilization by introducing the aggregation change factor to dy- namically change the particle population regeneration mechanism. To evaluate the proposed method, we extend cloudsim package to simulate the cloud environment, and compared with the double adaptive algorithra and the modified PSO algorithm ( ADPSO algo- rithm}. The simulation results show that the U-RPSO algorithm not only improves the efficiency of the task,but also reduces the cost of the task,and make full use of the physical resources in the cloud environment, which make the resource utilization rate maximized.
出处
《小型微型计算机系统》
CSCD
北大核心
2016年第6期1203-1206,共4页
Journal of Chinese Computer Systems
基金
湖北省教育厅科研计划指导性项目(B2015373)资助
关键词
虚拟技术
用户体验度
资源利用率
多目标粒子群算法
virtualtechnology
user experience
resource utilization
multi-objective particle swarm optimizati