一种基于改进樽海鞘算法的云仿真任务调度的研究

Research on Cloud Computing Task Scheduling Based on Salp Swarm Algothrim

云平台通常允许多个任务在云环境中同时执行,而任务调度是实现更好云计算性能的重要部分,其调度的效率直接影响到云平台计算资源利用率以及用户服务质量.针对云计算任务调度的核心寻求解的最优化问题,本文提出了一种混合算法,称为樽海鞘改进算法.此算法融合了反向学习原理扩大搜索空间,能够自适应的改变领导者的位置,并使得追随者根据几位领导者的位置更新自己,避免解陷入局部最优.本文采用CEC常用的23组测试函数进行测试,将结果与多个经典算法进行比较,证明了樽海鞘改进算法的优越性.同时在云仿真平台上进行模拟在云平台上进行任务调度的过程,通过与其他的几种算法的比较,证明了樽海鞘优化算法在任务调度方面应用的可行性,且有效缩短了云任务的完成时间,降低了完成成本.

Cloud platforms usually allow multiple tasks to be executed simultaneously in the cloud environment.Task scheduling is an important part of achieving better cloud computing performance,and its scheduling efficiency directly affects the utilization of cloud platform computing resources and user service quality.Aiming at the optimization problem of the core of cloud computing task scheduling,this paper proposes a hybrid algorithm,called LIL_Alarm Swarm Algorithm(LSA).This algorithm combines the reverse learning principle to expand the search space,and can adaptively change the position of leaders,and make followers update themselves according to the positions of several leaders,so as to avoid the solution falling into local optimum.In this paper,23 groups of test functions commonly used by CEC are used for testing,and the results are compared with several classical algorithms to prove the superiority of LSA algorithm.At the same time,the process of task scheduling on the cloud platform is simulated on the CloudSim platform.By comparing with other algorithms,it proves the feasibility of LSA optimization algorithm in task scheduling,and effectively shortens the completion time of cloud tasks and reduces the cost of completion.