Abstract
背景问题
Most algorithms developed for scheduling applications on global Grids focus on a single Quality of Service (QoS) parameter such as execution time, cost or total data transmission time. However, if we consider more than one QoS parameter (e.g. execution cost and time, which may be in conflict) then the problem becomes more challenging. To handle such scenarios, it is convenient to use heuristics rather than a deterministic algorithm.
为全局网格上的应用程序调度而开发的大多数算法都关注单个服务质量（QoS）参数，例如执行时间，成本或总数据传输时间。但是，如果我们考虑不止一个QoS参数（例如执行成本和时间，可能会发生冲突），则问题变得更具挑战性。 为了处理这种情况，使用启发法而不是确定性算法更加方有效。
方法
In this paper, we have proposed a workflow execution planning approach using Multiobjective Differential Evolution (MODE). Our goal was to generate a set of trade-off schedules according to two user specified QoS requirements (time and cost), which will offer more flexibility to users when estimating their QoS requirements.
在本文中，我们提出了一种使用多目标差分进化（MODE）的工作流执行计划方法。 我们的目标是根据两个用户指定的QoS要求（时间和成本）生成一组权衡时间表，这将为用户在估计其QoS要求时提供更大的灵活性。
实验验证
We have compared our results with a well-known baseline algorithm ‘Pareto-archived Evolutionary Strategy (PAES)’. Simulation results show that the modified MODE is able to find significantly better spread of compromise solutions compared with that of PAES.
我们将我们的结果与众所周知的基线算法'Pareto-archived Evolutionary Strategy（PAES）'进行了比较。 仿真结果表明，与PAES相比，改进的MODE能够找到明显更好的折衷解决方案。
Conclusion
方法
In this paper, we have proposed a workflow execution planning approach, which optimizes multiple objectives. The planner can generate a set of widespread alternative solutions if the optimization objectives are conflicted. Providing these alternative solutions can offer more flexibility to users to estimate their preferences and choose a desired workflow schedule based on their QoS requirements. Our MODE approach differs from others in the sense that we are dealing with real scheduling sequences rather than job/machine sequence to real number transformation as implemented in [7,8].
在本文中，我们提出了一种工作流执行计划方法，该方法可以优化多个目标。 如果优化目标存在冲突，规划人员可以生成一组广泛的替代解决方案。 提供这些替代解决方案可以为用户提供更大的灵活性，以估计他们的偏好并基于他们的QoS要求选择期望的工作流程表。 我们的MODE方法与其他方法的不同之处在于我们正在处理实际调度序列而不是作业/机器序列在[7,8]中实现的实数变换。
实验结果与结论
We have compared our results with PAES and the I−H indicator shows that our approach performs better than PAES. Moreover, our model is also free from extra computational overhead due to crowding distance sorting. In the case of candidate creation, we have only considered the Ulam distance and there may be further opportunities to investigate the algorithms performance with different string similarity metrics.
我们将结果与PAES进行了比较，I-H指标显示我们的方法比PAES表现更好。 此外，由于拥挤距离排序，我们的模型也没有额外的计算开销。 在候选人创建的情况下，我们只考虑了Ulam距离，并且可能有更多机会用不同的字符串相似性度量来研究算法性能。
畅想
Multiobjective optimization in Grid scheduling is not a matured field. It still requires a number of detailed benchmark problems to test every type of multiobjective scenario on Grid scheduling or real-life data to test an algorithm’s performance. There are also a limited number of studies that have considered the flow-shop/job-shop scheduling problem using DE, however, multiobjective scheduling poses additional challenges. Many of the widely used existing workflow scheduling algorithm only attempt to minimize either execution time or execution cost. However, additional objectives must be considered when scheduling workflows on utility Grids.
网格调度中的多目标优化不是一个成熟的领域。 它仍然需要一些详细的基准测试问题或实际数据的每种类型的多目标场景网格调度来测试算法的性能。 还有一些研究使用DE考虑了流水车间/车间调度问题，但是，多目标调度带来了额外的挑战。 许多广泛使用的现有工作流调度算法仅尝试最小化执行时间或执行成本。 但是，在公用事业网格上安排工作流时，必须考虑其他目标。
