基于天牛须遗传混合算法的大规模云任务调度

Large-Scale Cloud Task Scheduling Based on Beetle Antennae Search Algorithm Hybrid Genetic Algorithm

针对目前云计算环境下大规模任务调度效率低下的问题,提出一种基于天牛须搜索算法与遗传算法混合的优化算法.为改善遗传算法优化结果不稳定和局部搜索能力差的问题,采用新型交叉方式,动态改变变异概率,拓展天牛搜索方向,并采用精英解保留策略.最后基于CloudSim仿真平台,在不同任务规模调度情况下,将混合算法和其他3种算法在相同环境下进行仿真调度.实验结果表明:该混合算法比其他算法得到更优质解,相比于遗传算法,小规模任务调度优化结果提高7.87%,大规模任务调度提高30.23%,是一种高效的云任务调度优化算法.

Aiming at the low efficiency of large-scale cloud task scheduling in the current cloud computing environment,an optimization algorithm based on beetle antennae search and genetic algorithm is proposed in this article.In order to improve the unstable optimization results and poor local search ability of genetic algorithm,a new crossover method is adopted,and the mutation probability is changed dynamically.Then the search direction of beetle is expanded,and the elite solution retention strategy is adopted.Finally,based on CloudSim,the hybrid algorithm and the other three algorithms are simulated and scheduled in the same environment in the case of different task scale scheduling.The experimental results show that the hybrid algorithm obtains better solutions than other algorithms.Compared with genetic algorithm,the optimization result of small-scale task scheduling is improved by 7.87%and that of large-scale task scheduling is improved by 30.23%,it proves an efficient cloud task scheduling optimization algorithm.