基于萤火虫算法和动态优先级的多QoS云工作流调度

Multi-QoS Cloud Workflow Scheduling Based on Firefly Algorithm and Dynamic Priorities

为了提高用户的满意程度、降低运营成本,考虑时间、费用和可靠性3个重要的服务质量因素,针对时间和可靠性双重约束下费用最小化的云工作流调度问题,提出基于萤火虫算法和动态优先级的最优调度方案。结合云工作流调度问题的特点,重新定义了萤火虫算法中的位置、距离以及位置更新方式,同时对于每一种调度方案,采用动态优先级算法确定任务顺序,以减少工作流完成时间。在WorkflowSim平台上进行模拟调度仿真实验,证明了该方法在收敛速度和最优值方面均优于传统的云工作流调度算法。

To improve the user satisfactory and reduce the operating costs, three important attributes of Quality of Service （QoS） which included time, costs and reliability were taken into account. To solve the scheduling problem of cloud workflow with minimized cost and constrains between time and reliability, an optimal scheduling approach based on firefly algorithm and dynamic priorities was proposed. The position, distance and updating mode of position in firefly algorithm were redefined and the priorities were dynamically set for task order to reduce the overall completion time. The experimental results conducted on WorkflowSim demonstrated that the proposed approach was superior to the traditional cloud workflow scheduling algorithm with regard to both the convergence speed and the obtained optimal value.