中文摘要：

针对云计算环境下资源调度未能全面考虑QoS需求的问题，提出了一种多维QoS约束的调度方法。采用公式化描述和性能分析对多维QDS进行建模，在优化方法中，引入系统指标，通过偏好因子改变权重，结合主元和客元的合成以及公平性评判，实现了对多目标优化问题的非劣最优解的锁定。实验结果表明，该方法充分考虑了用户多维QoS需求，在系统开销相当的情况下，负载均衡程度明显提高，且平均提高了17％，资源分配能更好的符合用户满意度要求。

英文摘要：

Aiming at the problem of incomplete consideration of QoS requirements in online cloud scheduling, a scheduling approach with multidimensional QoS constraints is put forward. The formal definition and the performance analysis is employed in the modeling of multidimensional QoS. In order to fix Pareto optimal solution, system index is introduced by changing the weights in preference factor based on the synthesis of principal component and factor as well as fairness judgement. The experiment result shows that the approach gives sufficient consideration of multidimensional user QoS requirements. With the same system costs, the extent of load balance is significantly improved by an average of 17%. It can help allocate resources to better serve user satisfaction.