In order to improve the efficiency of computing of intelligence algorithm in Map-Reduce job scheduling problems,this paper proposed an improved shuffled frog leaping job scheduling algorithm.Specifically,the algorithm designed the coding plan and evolutional operators,targeted in solving Map-Reduce job scheduling problems with frog methodology;and improved frog methodology,utilizing diversity of population to increase inversion mutational strategy,and improved the convergence capability of the algorithm.In simulation experiment,the improved frog strategy has more capability in convergence,total task time and average task time compared with the traditional frog methodology and up-to-date intelligence algorithm in solving MapReduce job scheduling problems,and thus a practical approach in Map-Reduce scheduling.%为提高智能算法在Map-Reduce作业调度问题中的求解效率,提出一种基于改进蛙跳策略的调度算法.针对蛙跳策略在Map-Reduce作业调度中的应用,算法具体设计了编码方案和进化算子；同时,为提高算法收敛性能,对蛙跳策略进行改进:结合种群多样性指标增加逆转变异操作.仿真实验结果表明,提出的改进蛙跳策略在Map-Reduce作业调度问题求解中,收敛性能、作业总完成时间和平均完成时间三个方面均优于基本蛙跳策略和已有的智能调度算法,是一种实用的Map-Reduce作业调度方案.
展开▼
