Vague集多QoS属性ER测量方法及网格任务调度

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，用证据理论盼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（VSFM QoS-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。

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

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.