基于多属性QoS约束的网格资源调度算法研究

摘要

本文以Min-Min调度算法为基础,以提高调度过程中的用户满意度为目的,提出了改进的Min-Min算法DQ(double QoS).在DQ算法的基础上,采用进一步细化QoS参数和划分任务权值的方法,设计了多属性QoS约束的调度算法(multi-QoS constraints scheduling algorithm,MQCSA).该算法通过选取任务的完成期限和网络带宽属性以及完成任务所需的CPU参数,有效地解决了不能预期完成的任务占用网格资源的问题.采用GridSim工具包模拟了相关的调度算法,验证了DQ算法和MQCSA的效率.%To improve the utilization ratio of resources and users' satisfaction degree of grid services, the algorithm DQ based on improved Min-Min algorithm was proposed in this paper. On the basis of DQ algorithm, applying the method of refining QoS parameter and classifying task weight value, multi-QoS constraints MQCSA was proposed. By choosing deadline of gridlets, bandwidth of network and CPU parameters, the algorithm could resolve the problem of occupying resources thai the tasks cannot complete on schedule. It was proved that MQCSA and DQ algorithm are superior in solving mulli-QoS constraints problems by comparing and analyzing simulation data.