基于烟花算法的云计算任务调度研究  

Research on Cloud Computing Task Scheduling Based on Firework Algorithm

作　　者：孙凤杰 王克俭[1] 何振学 高万豪 SUN Feng-jie;WANG Ke-jian;HE Zhen-xue;GAO Wan-hao(College of Information Science and Technology Hebei Agricultural University,Baoding Hebei 071001,China;College of Control and Mechanical Engineering Tianjin Chengjian University,Tianjin 300384,China)

机构地区：[1]河北农业大学信息科学与技术学院,河北保定071001 [2]天津城建大学控制与机械工程学院,天津300384

出　　处：《计算机仿真》2022年第3期340-343,443,共5页Computer Simulation

基　　金：河北省可信信息系统重点实验室开放研究课题(2018KFKT003);河北省教育厅重点研究项目(ZD2016158);河北农业大学引进人才科研专项(YJ201829)。

摘　　要：通常云计算中需要处理大量的任务,云任务调度策略在云计算效率方面起着重要的作用,并且是一个热点研究方向。现提出一种基于烟花算法的任务调度算法。算法以任务完成时间和完成成本为优化目标。根据爆炸火花,适应度值较大的烟花产生较多的火花,反之,产生较少的火花,实现算法的局部和全局搜索;根据高斯变异火花,增加了种群的多样性,有利于将好的个体传递到下一代烟花种群中。最后,通过CloudSim平台进行仿真,在任务完成时间和成本指标下,将所提算法与GA、PSO和TCGA作对比,结果表明,所提算法有效缩短了云任务的完成时间,降低了完成成本,并且优于其它三种对比算法。Usually cloud computing needs to process a large number of tasks.Cloud task scheduling strategy plays an important role in cloud computing efficiency and is a hot research direction.Based on the firework algorithm,this paper proposed a task scheduling algorithm which took the task completion time and completion cost as the optimization goals.According to the explosion sparks,fireworks with larger fitness values generated more sparks,and conversely,generates less sparks,and realized the local and global search of the algorithm;According to Gaussian mutation sparks,the diversity of the population was increased,which is conducive to passing good individuals to the next generation of fireworks populaion.Finally,through the CloudSim platform for simulation,under the task completion time and cost indicators,the algorithm of this paper was compared with GA,PSO and TCGA,and the results show that the algorithm of this paper effectively shortens the completion time of the cloud task,reduces the completion cost,and optimizes for the other three comparison algorithms.

关 键 词：云计算 任务调度 烟花算法 时间 成本 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]