基于改进粒子群算法的虚拟机放置算法

摘要

云计算通过使用虚拟机技术提高了数据中心的资源利用率.虚拟机放置算法作为云计算的关键技术,具有重要研究意义.现有虚拟机放置算法往往只关注成本控制和云资源使用率,忽略了负载均衡对系统性能的影响.针对该问题,本文在标准的粒子群优化算法基础上进行改进,首先设计多目标函数时引入负载不均衡度概念,然后通过系统实时负载随机生成初始化种群,并在算法中引入分组思想,通过对初始种群进行随机分组,避免算法陷入早熟现象.通过 CloudSim模拟平台进行仿真实验,表明改进后的算法更利于云数据中心进入负载均衡状态,并有较高的资源利用率.%Cloud computing through the use of virtual machine technology has greatly enhanced the utilization of cloud resource in the data center.As the key technology of cloud computing,virtual machine placement algorithm has important significance.Existing virtual machine placement algorithms tend to focus on cost control and utilization of cloud resources,ignoring the impact of load balancing on system performance.To solve this problem,this paper im-proved the standard particle swarm optimization algorithm.Firstly,this paper introduces the concept of load imbalance when designing multi-objective function,and then randomly generates initial population through the real-time load,and the introduction of grouping idea,avoiding falling into premature.By CloudSim simulation experiment platform,it shows that the improved algorithm is more conducive to cloud data centers into balance,and has a higher resource utilization.