中文摘要：

为解决现有的虚拟机（VM）初始放置目标较为片面,仅考虑1个或者2个方面的问题,该文提出了1种面向多目标优化的VM初始放置方法。综合考虑了资源利用率、功率以及温度3方面因素。基于改进的分组遗传算法生成候选的VM放置方案。采用多目标模糊评估方法筛选出最佳放置方案。仿真实验结果表明,该文方法可以减少约44%的资源浪费、降低3 k W服务器运行功率。

英文摘要：

To solve the problem of existing virtual machine placement methods that the initial placement target is one-sided and only focuses on one or two optimization objects ,a virtual machine initial placement method for multi-objective optimization is proposed here .Resource usage rate , system power and temperature are considered synthetically .Candidates of virtual machine placement solution are got based on an improved group genetic algorithm .The best virtual machine placement solution is selected by a multi-object fuzzy assessment algorithm .The simulation experiment results show that the proposed method can reduce the wasting of resources by 44% and server operation power by 3 kW.