Vague集多QoS属性ER测量方法及网格任务调度

ER-Based Vague Multi-QoS Measurement and Grid Task Scheduling

针对网格环境中的不确定性,提出一种基于ER(Evidential Reasoning)的Vague集多QoS(Quality of Service)属性测量方法。该方法用Vague集表示用户及资源的模糊QoS属性,用证据理论的ER算法将分层的QoS属性融合为综合效用值。通过算例分析,证明了该方法的可行性和正确性。进一步将该测量方法结合到网格任务调度,提出一种网格任务调度算法(VSFMQoS-Guided-TS),通过仿真结果表明,VSFMQoS-Guided-TS在保留QoS不确定性的情况下,减小了调度时间跨度,提高了调度效率。

To cope with the uncertainty of Grid environment, an ER （Evidential Reasoning） based Vague multi,QoS （Quality of Service） measuring approach is proposed. The approach utilizes Vague set to denote the QoS of users or and resources, and utilizes D-S theory based ER algorithm to integrate the layered multi-QoS into a single utility value. The calculation example proves the feasibility and validity of the approach： Further- more, the approach is applied to grid task scheduling, and a new grid task scheduling algorithm （VSFM- QoS-Guided-TS） is presented. Experimental results demonstrate that the algorithm does not only reserve the uncertainty of QoS but also reduces the scheduling makespan and improves the scheduling efficiency.