基于QK-Mean聚类的云任务调度算法

A Cloud Task Scheduling Algorithm Based on QK-Mean Clustering

在运用聚类的思想对云资源进行分类时,资源属性的信息熵能反映资源属性在聚类过程中的重要程度。将信息熵的思想运用到K-mean聚类中,提出基于QK-mean聚类的任务调度算法,即对云资源按照属性的重要程度逐一进行K-mean聚类,把任务的调度过程转化成在资源K叉树中查找合适的"叶子"节点的过程。实验证明,该任务调度算法能够有效提高云任务调度的效率。

The comentropy can reflect the importance of resource properties in clustering process, when cloud re- sources are classified with the idea of clustering. The idea of comentropy was applied to K-mean clustering to propose task scheduling algorithm based on QK-mean clustering, which the cloud resources was clustered according to the im- portance of the properties. It transformed the task scheduling process into the process of finding a suitable ＂leaf＂ in resources K-tree. The experimental results showed that the scheduling algorithm can improve effectively the efficiency of cloud task scheduling.