Dynamic deployment of virtual machines in cloud computing using multi-objective optimization
Dynamic deployment of virtual machines in cloud computing using multi-objective optimization
复制标题
使用多目标优化在云计算中动态部署虚拟机
DOI：
10.1007/s00500-014-1406-6
复制
发表时间：
2015-08-01
期刊：
影响因子：
4.1
通讯作者：
Yu, Jian-Ping
中科院分区：
文献类型：
--
作者：
Xu, Bo;Peng, Zhiping;Yu, Jian-Ping
云计算被认为是第五种实用服务，是下一代计算。计算资源可以根据用户的需求和偏好进行动态分配虚拟机部署在云计算中具有重要作用，旨在减少周转时间并提高资源利用率。从本质上讲，虚拟机的部署是一个多目标决策问题，必须考虑关键因素。也就是说，我们需要优化资源使用和迁移时间。本文提出了虚拟机动态部署的多目标综合评价模型。然后，我们使用一种改进的多目标粒子群优化算法（IMOPSO）来解决这个问题。利用CloudSim工具包设计了两个仿真实验：第一个实验结果表明，改进算法与传统的单目标PSO和QPSO算法相比，具有可行性和高效性;第二个实验结果表明，IMOPSO算法能够有效地搜索，保持种群多样性，快速收敛到Pareto最优解而不失去稳定性。所得到的Pareto最优解集具有比比较方法更好的收敛性和分布性。
Cloud computing is regarded as the fifth utility service and is the next generation of computation. The computing resources can be dynamically allocated according to consumer requirements and preferences Virtual machine deployment has an important role in cloud computing, and aims to reduce turnaround times and improve resource use. In essence, the deployment of virtual machines is a multi-objective decision problem that must consider key factors. That is, we need to optimize the resource use and migration times. In this paper, we propose the multi-objective comprehensive evaluation model for the dynamic deployment of virtual machines. We then use an improved multi-objective particle swarm optimization (IMOPSO) to solve the problem. We have designed two simulation experiments using the CloudSim toolkit: the first experimental results show that on comparison of our improved algorithm with the traditional single-objective algorithms PSO and QPSO, our method is feasible and efficient; the second experimental results show that IMOPSO can search effectively, maintain population diversity, and quickly converge to the Pareto optimal solution without losing stability. The obtained Pareto optimal solution set has a better convergence and distribution than a comparative method.