云计算中基于NSGA Ⅱ的虚拟资源调度算法

摘要

The physical resources are usually abstracted the identical virtual resources in cloud computing, it is a basic and com plex problem on how to schedule virtual resources to physical resources in cloud computing. The scheduling of virtual resources is abstracted by a mathematic model and its intractability is proved, then this model is changed to be a multi-goals optimization problem based on system loading balancing, the method of solving this problem employing the advanced non-dominated sorting genetic algorithm II (NSGA II) is proposed. Contrasted with the concrete implementation, the abstracted model is more suitable to describe the scheduling problems in cloud computing. Simulation model verifies the effectiveness of the model and the possibility of employing NSGA II to solving the scheduling problem by experiment, compared with the random algorithm, static algorithm and matching algorithm, NSGA II obtain better results.%云计算环境下将物理资源抽象为同一的虚拟资源,如何将虚拟资源调度到物理资源上是云计算中一个基本且复杂的问题.对虚拟资源的调度进行建模并证明其难解性,将该模型的求解转化以系统负载均衡为优化目标的多目标优化问题,提出采用改进的基于非支配排序的遗传算法(NSGA Ⅱ)来求解该问题.与针对具体环境的调度算法相比,抽象的模型更能代表典型的云计算环境中的虚拟资源调度问题.对提出模型进行了仿真,实验结果表明了该模型的有效性和NSGA Ⅱ算法求解该问题的可行性,同时对比随机算法、静态算法和排序匹配调度算法,NSGA Ⅱ算法优于其他算法.