IDEA融合Taguchi方法的云计算任务调度优化算法  

A Task Scheduling Optimization Algorithm Based on Fusion of IDEA and Taguchi in Cloud Computing

作　　者：李敬伟 孙波 

机构地区：[1]河南工学院计算机科学与技术系,河南新乡453002

出　　处：《控制工程》2017年第2期458-466,共9页Control Engineering of China

基　　金：河南省高等学校重点科研项目(16B520009);河南省高等学校重点科研项目(15A520062)

摘　　要：为了优化云计算环境下的任务调度,提出了一种基于改进差分进化算法融合Taguchi方法的任务调度优化算法。首先,利用基于子任务资源的技术进行编码调度;然后,利用掩膜突变算子代替DEA中传统数学算子,从而生成突变个体并在交互中改变子任务和重组资源;最后,将带有改进算子的DEA与带有正交掩膜的Taguchi方法合并实现联合优化。实例分析结果表明,相比DEA和NSGA-II,该算法在优化任务调度和资源分配方面具有更好的性能,并且减少了总完工时间和调度成本。To optimize the task scheduling and the resource allocation in cloud computing, an optimization algorithm based on the improved differential evolution algorithm （IDEA） and Taguehi method is proposed. Firstly, a subtask-resource-based technique is used to encode a schedule. Then, a mask mutation operator instead of a traditional mathematical operator in DEA is utilized to generate a mutated individual and the subtasks are changed in mutation and the resources are recombined in crossover. Finally, the method in this paper is proposed by combining the DEA with the improved operators and the Taguchi method with the orthogonal mask. In the study, the IDEA shows its effectiveness to optimize the task scheduling and resource allocation compared with both the DE A and the NSGA-II, and it has reduced the makespan and cost significantly.

关 键 词：云计算 改进差分进化算法 任务调度 突变算子 Taguchi方法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]