为提高云计算任务调度的效率,减少系统执行任务的最大完工时间以及成本,本文提出一种改进的人工鱼群任务调度算法(improved artificial fish swarm algorithm,IAFSA).首先,将反向学习策略应用于种群初始化和鱼群的行为选择中,以提高改进人工鱼群算法在迭代中的收敛速度和种群多样性.其次,将自适应全局-局部记忆机制引入到标准AFSA算法的觅食行为中,以进一步提高勘探能力.最后,增加了基于平均适应度的行为选择机制,以提供更合理的行为选择,减少算法的复杂性.通过使用CloudSim 平台进行实验验证,分别测试在不同任务规模下IAFSA的算法效能.实验结果表明,改进人工鱼群算法在降低系统任务最大完工时间和成本上均表现出了显著的优势.
Abstract
In order to improve the efficiency of cloud computing task scheduling and reduce the makespan and cost of tasks,this paper proposes an improved artificial fish swarm task scheduling algorithm(IAFSA).Firstly,the opposition-based learning strategy was applied to the population initialization and the behavior selection of the fish swarm to improve the convergence speed and population diversity of the improved artificial fish swarm algorithm in iterations.Secondly,the adaptive global-local memory mechanism was introduced into the foraging behavior of the standard AFSA algorithm to further improve the exploration ability.Finally,an action selection mechanism based on average fitness was added to provide more reasonable action selection and reduce the complexity of the algorithm.By using CloudSim platform for experimental verification,the algorithm efficiency of IAFSA under different task scales was tested respectively.The experimental results show that the improved artificial fish swarm algorithm has significant advantages in reducing the maximum completion time and cost of the system task.
维普
万方数据