CAJ | 학술논문

在云环境中,如何将大量的虚拟机调度到物理节点上是一个基本且复杂的问题。文中首先对虚拟机的调度建立装箱问题模型,将该模型的求解转化一个多目标优化问题,目标分别为负载均衡、提高任务执行效率和降低能耗；接着对基于非支配排序的遗传算法( Non-dominated Sorting Genetic Algorithm,NSGA II)进行改进,利用回溯法中的剪枝函数确定最优初始种群,引入正态分布密度函数限制优秀精英。仿真结果表明,基于改进NSGA II的虚拟机调度算法在任务执行时间、负载均衡和能量消耗三个方面优于其他一些常用算法。
재운배경중,여하장대량적허의궤조도도물리절점상시일개기본차복잡적문제。문중수선대허의궤적조도건립장상문제모형,장해모형적구해전화일개다목표우화문제,목표분별위부재균형、제고임무집행효솔화강저능모；접착대기우비지배배서적유전산법( Non-dominated Sorting Genetic Algorithm,NSGA II)진행개진,이용회소법중적전지함수학정최우초시충군,인입정태분포밀도함수한제우수정영。방진결과표명,기우개진NSGA II적허의궤조도산법재임무집행시간、부재균형화능량소모삼개방면우우기타일사상용산법。
In cloud environment,how to schedule the large number of virtual machines to the physical nodes is a fundamental and difficult problem. Firstly,establish bin packing problem model based on virtual machines scheduling and solves the model through transforming it into a multi-objective optimization problem. The objectives respectively are load-balancing,to improve task-execution efficiency and to reduce energy consumption. Then,non-dominated sorting genetic algorithm is improved. The pruning function in the backtracking is used to confirm the optimal initial population. The normal distribution density function is introduced to restrict elite. The results of simulation show that the virtual machines scheduling algorithm based on improved NSGA II on the aspects of task execution time,load-balancing and energy consumption is better than other commonly used algorithms.