一种成本和时间均衡的网格资源分类优化调度策略  

Grid Resource Classified Optimization Scheduling Strategy for Cost and Time Balancing

作　　者：姚念民[1] 李昂[1,2] 陈金忠[1] 况奇伟 

机构地区：[1]哈尔滨工程大学计算机科学与技术学院 [2]解放军92474部队

出　　处：《小型微型计算机系统》2012年第12期2653-2657,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61073047)资助;高效能服务器和存储技术国家重点实验室开放基金项目(2009HSSA08)资助;中央高校基本科研业务费专项资金(HEUCFT1007;HEUCF100607)资助

摘　　要：针对时间和成本约束的网格资源调度问题,提出一种基于MinCTT算法的时间和成本均衡的网格资源分类优化调度算法.该算法综合考虑任务完成时间和执行成本两个QoS因素,由一个成本比值和时间比值的联合均衡值来综合衡量任务在资源上的完成时间和执行成本开销,根据任务估计平均价格,对资源进行分类调度.实验结果表明,该调度算法具有较好的调度性能,能有效的减少任务总的完成时间和执行成本,均衡因子的改变对该算法的调度性能影响较小,选择合适的均衡因子能实现优的调度.Concerning the time and cost constraint for grid resource scheduling problem, a grid resource classified optimization schedu- ling algorithm for cost and time balancing which is based on MinCTT scheduling algorithm is proposed. The algorithm synthetically considers two kinds. QoS factors： the makespan and execution cost of tasks. To comprehensively measure the makespan and execution cost of the task on the resource by a joint balancing value that form of a cost ratio and a time ratio, according to the estimated average price of tasks, to assort resources for scheduling, tasks. Experimental results show that our scheduling algorithm has better scheduling performance and can effectively reduce the overall makespan and execution cost of tasks. The change of balancing factors has a small impact on the performance of our scheduling algorithm, it can achieve good scheduling when proper balancing factors are choosed.

关 键 词：网格 调度 完成时间 执行成本 均衡因子 GRIDSIM 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]