摘要： 实现用户的服务质量(QoS)是网格计算中力求达到的重要目标，网格资源的分布性、异构性、动态性等特征使网格环境下以服务质量为指导的资源调度成为一个复杂的问题，尤其是在用户的任务具有多种QoS属性的情况下。该文利用经济模型研究网格QoS控制的资源分配问题。以效用最大化为目标通过综合效用函数量化服务质量，设计了在时间和费用受限情况下对任务进行分类的优化调度算法，该调度算法满足用户多QoS属性。仿真实验显示了该算法的有效性。

关键词: 多QoS属性, 资源调度, 经济网格, 分类优化

Abstract: It is because the resources are geographic distributed, heterogeneous and dynamic in computational grid environments that makes QoS guided scheduling a complex and challenging problem, especially when the tasks have multiple QoS needs. The economic model is applied to solve the resource allocation problem controlled by QoS in grid. Integration function is used to quantitative QoS as an object to achieve high system utilization. A classified optimization scheduling algorithm for a set of independent tasks under the limitation of time and cost is proposed which can satisfy the multi-QoS attributes effectively. The results of the simulation in virtual Grid environment show that the algorithm can well tradeoff the actual executing cost and its budget in grid economy model.

Key words: multi-QoS attributes, resource scheduling, economical grid, classified optimization

中图分类号: 

• TP393