网格环境下基于多QoS约束的任务调度算法研究

摘要

针对网格任务调度过程中的QoS （Quality of Service）约束可扩展性不足、资源利用率低等问题，提出一种多QoS约束的Min_Min任务调度算法。该算法首先计算待调度任务QoS与每个资源QoS的匹配值，并根据匹配值的大小确定任务分配的先后顺序，然后再依次对已排序的任务选取具有最小最早完成时间所对应的资源进行分配。在保证QoS匹配质量的同时，也注重资源节点的利用率。仿真实验结果表明，该算法在综合性能上明显优于Min_Min算法。%To solve the problem of lack of scalability and low rate of resource utilization in QoS constrained grid task scheduling process, in this paper, a Min_Min task scheduling algorithm with multiple QoS constrains is put forward. Firstly, the similar matching values of the tasks’ QoS and the resources’ QoS is computed by the algorithm, and the task allocation sequence according to the matching values is determined.And then in turn for the ordered task, the re-source which has the smallest earliest finish time is allocated.Both of the quality of the matching values and the utiliza-tion rate of the resources would be concerned among the scheduling process.The results of experimental simulation show that the comprehensive performance of the new algorithm is better than Min_Min algorithm.