基于路径平衡的工作流费用优化方法

针对效用网格下截止期约束的工作流费用优化问题,提出了路径平衡(path balance,简称 PB)算法,对工作流中各路径长度进行调整,并提出基于路径平衡的费用优化(path balance based cost optimization,简称PBCO)算法。 PBCO 基于 PB 的计算结果设置初始约束时间,充分利用了工作流的费用优化空间。同时,采用逆向分层策略对任务进行分层,并根据各层任务数按比例分配冗余时间,有效地增大了多数任务的费用优化空间,进一步改善了工作流的费用优化效果。实验结果表明,PBCO比另外几种著名算法(如DET,DBL等)改进了约35%。

Path Balance Based Heuristics for Cost Optimization in Workflow Scheduling

In order to address the trade-off problems between time and cost in the grid workflow scheduling with deadline constraints, this study proposes a new algorithm named path balance (PB) to adjust the length for each path in workflow and develop a heuristic referred as path balance based cost optimization (PBCO). PBCO makes full use of the workflow cost optimization space by setting a deadline for each task, based on the results of PB, and enlarges the optimization space for multitude tasks by distributing the redundancy time, based on the quantities of tasks at each level, this is divided by using the bottom level strategy. According to the experimental results, the average execution cost of PBCO decreases by about 35% relative to other famous heuristics, such as DET and DBL.