有向无环图描述的工作流时间费用优化问题是计算网格中的基本难题之一。针对该问题提出一种网格工作流的动态贪心调度算法(DGS-GW)。按用户提交的截止时间和偏好因子建立数学模型,利用DGS-GW求解该模型的次优解。实验结果证明了该算法的有效性。
Optimization problem of workflow time-cost described by Directed Acyclic Graph(DAG) is a fundamental intractable problem in computational grid. Aiming at this problem, this paper proposes a Dynamic-Greedy Schedule algorithm for Grid Workflow(DGS,GW). According to the deadline and preference factor submitted by users, it establishes a mathematical model and uses DGS-GW to solve it. Experimental results show that the algorithm is effective.