摘要
IaaS(infrastructure as a service,基础设施即服务)模式云训练是一种以云计算为基础的新型装备模拟训练.云训练中,虚拟机的优化放置是提高资源利用效率、降低运行时资源调度工作量的基础.阐述了云训练的内涵,并对云训练中的虚拟机放置进行了数学描述.提出了一种改进的免疫克隆优化算法(MICOA),采用反向优化算法对初始抗体进行优化,通过变异概率与范围的自适应控制,保证算法演化初期抗体群的多样性与搜索空间的完备性,以及演化后期的局部寻优与最优解质量.引入抗体-抗体亲和度筛选最优抗体,保证抗体群的多样性.通过对虚拟机放置进行仿真实验,表明该方法可以有效提升资源利用率,实现系统综合优化目标.
IaaS （infrastructure as a service） mode cloud training is a new kind of training mode based on cloud computing.In cloud training, the optimal placement of virtual machines （VM） is the base of improving resource utilization and reducing workload of lower level resource scheduling. This paper first describes the meaning of cloud training and give the mathematical description of the problem. Then Modified Immune Clone Optimization Algorithm （MICOA） is proposed.Inversing optimization algorithm is adopted to process the randomly generated initial antibodies.Then through self adjusting of the probability and range of antibody mutation, the diversity of antibody population and the completeness of solution space are achieved. The antibody-antibody affiliation is introduced to realize screening of optimal antibodies.Experimental results show that the method can effectively improve the resource utilization and realize comprehensive optimization target.
出处
《军械工程学院学报》
2015年第2期54-58,共5页
Journal of Ordnance Engineering College
基金
装备预研基金(140A04030213JB34001)
关键词
云训练
虚拟机放置
反向优化
变异自适应控制
免疫克隆
cloud training
VM placement
inversing optimization
self adjusting of mutation
immune clone
