多QoS约束网格作业调度问题的多目标演化算法

针对网格计算中的多QoS约束网格作业调度问题，以独立作业为研究对象，将其规约为多目标组合最优化问题．通过深入剖析多目标最优化理论及其演化算法，结合网格作业调度自然特征，提出了一种解决多QoS约束网格作业调度问题的多目标演化算法．该算法求解多个QoS维度效用函数指标的非劣解集，尝试解决多管理域间网格用户、资源管理者等网格实体的多目标协同问题．仿真结果表明，在时间维度、可靠性维度、安全性维度QoS效用值等用户级QoS指标，以及丢弃作业数等系统级指标方面该算法与QoS-Min-min和QoS-Sufferage等同类算法相比具有较好的综合性能．

A Multiobjective Evolutionary Algorithm for Grid Job Scheduling of Multi-QoS Constraints

Grid scheduling and resource management potentially involve the interaction of many human players such as end users and resource administrators. Such players result in different, often contradictory, criteria and make the process of mapping jobs to resources difficult or even impossible. Focusing on the independent jobs, the multiobjective job scheduling problem of multi-QoS constraints is proposed and transformed to the general multiobjective combinatorial problem. An advanced evolutionary algorithm is put forward to solve multiobjective grid job scheduling. The evolutionary technique is used to find the non- dominated set of solutions and distribute them uniformly in the Pareto front so that the best compromise scheduling solution can be found. It is shown via simulation that the algorithm performs better than the QoS-Min-min and QoS-Sufferage in the user-QoS performances such as time-dlmension, reliability-dimension, security dimension QoS utilization and the system-QoS performance such as dropped job numbers.