CAJ | 학술논문

云计算环境下虚拟机的部署是有关资源管理、服务质量等方面的综合性问题，是一个多目标的NP-hard问题。根据弹性云环境下虚拟机部署的特点，利用遗传算法解决多目标组合优化问题的优点，提出一种基于改进遗传算法的虚拟机部署算法，以减少物理计算机开销、提高资源利用率、满足用户SLA为优化目标，将云计算环境下的资源管理和服务质量需求转化为算法可识别的信息，通过改进遗传算法，使之适应于静态和动态部署的要求。实验结果分析表明，该算法在满足多目标的条件下得到较优的解，为虚拟机到物理节点的映射提供了较好的解决方案。
운계산배경하허의궤적부서시유관자원관리、복무질량등방면적종합성문제，시일개다목표적NP-hard문제。근거탄성운배경하허의궤부서적특점，이용유전산법해결다목표조합우화문제적우점，제출일충기우개진유전산법적허의궤부서산법，이감소물리계산궤개소、제고자원이용솔、만족용호SLA위우화목표，장운계산배경하적자원관리화복무질량수구전화위산법가식별적신식，통과개진유전산법，사지괄응우정태화동태부서적요구。실험결과분석표명，해산법재만족다목표적조건하득도교우적해，위허의궤도물리절점적영사제공료교호적해결방안。
Virtual machine deployment in cloud computing environment is a comprehensive problem that involves resource management, quality of service and so on, which is proved to be a NP-hard problem. According to the characteristics of virtual machine deployment under elastic cloud compu-ting environment, a virtual machine deployment algorithm based on improved genetic algorithm with advantages of solving multi-objective optimization problem is proposed. With the optimization objec-tives including reducing physical computer costs, improving resource utilization and satisfying users’ SLA, the algorithm transfers the requirements of resource management and quality of service into i-dentifiable information, and uses the improved genetic algorithm to adapt to both static and dynamic deployment. Experimental results show that the presented algorithm can get better optimization re-sults to meet multiple objectives, which provides a better solution for mapping between virtual ma-chine and physical node.