分布式计算中基于A-star的工作流调度改进算法研究  

Research of the A-star-based workflow scheduling advanced algorithm for distributed computing

作　　者：李坤[1] 蒋莉莉[2] 

机构地区：[1]中国石油大学计算机与通信工程学院,山东青岛266555 [2]东营职业学院计算机系,山东东营257091

出　　处：《计算机工程与科学》2013年第3期38-42,共5页Computer Engineering & Science

摘　　要：异构分布式系统中,若考虑每个任务的中间数据传输时间和计算时间,工作流调度问题就变得很难解决。论文在研究基于A-star的数据感知算法的基础上,通过在计算节点上进行任务执行和数据部署的重叠操作来实现最优调度。模拟结果显示,在大多数情况下,改进后的算法在性能和时间效率上要优于现有算法,明显降低工作流程周转时间。此外,也通过扩展所提算法来解决流程联合调度问题。The due to both the int workflow ermediate The paper has a study mal scheduling which sites. The simulation work in performance extend the algorithm of the scheduling problem in data transfer time and data-aware workflow s is through the overlappin results show that, in most and efficiency, and to solve the process g o cas heterogeneous distributed systems is hard to solve the computation time for each task being considered. cheduling algorithm based on A-star, to achieve opti- f task execution and data deployment on computing es,the improved algorithm is superior to the existing significantly reduces the turnaround time. In addition, we also co-scheduling problem

关 键 词：工作流调度 大规模科学计算 最优调度 周转时间 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]