A cloud service workflow scheduling algorithm based on novelty ranking and multiple quality of service (QoS) was proposed, aiming at the problem that the optimal scheduling based on user cost and system utilization was not considered in the existing researches.Frequency of the tasks performed by the resource node, waiting time and execution time of the resource node were added into the recommendation model.The simulated annealing algorithm was used to train the recommendation model, and the priority factor was calculated.The scheduler performed the scheduling and updated it according to the priority factor table.Simulation results show that the proposed algorithm is better than the Q-learning algorithm in terms of task execution time, and the combined index of user cost and system utilization is better than that of the Q-learning algorithm on CloudSim platform.%针对现有研究未能综合考虑以用户成本和系统利用率为目标进行优化调度的问题,提出基于新颖性排名和多服务质量(QoS)目标的云工作流调度算法.将资源节点执行任务的频度、任务的等待时间和执行时间作为因子加入推荐模型;使用模拟退火算法训练得到推荐模型,计算出优先级因子;调度器根据优先级因子表进行调度并对其进行更新.在CloudSim平台上进行模拟调度仿真实验,结果证明:所提出算法的任务执行时间优于Q值学习(Q-learning)算法,且用户成本和系统使用率的综合指标更好.
展开▼