基于改进CRO算法的Pareto多目标科学工作流调度算法  

Research on Pareto multi-objective scientific workflow scheduling algorithm based on CRO algorithm

作　　者：马敬敬 阎朝坤[1] MA Jingjing;YAN Chaokun(College of Computer and Information Engineering,Henan University,Kaifeng 475000,China)

机构地区：[1]河南大学计算机与信息工程学院,河南开封475000

出　　处：《周口师范学院学报》2018年第2期117-121,共5页Journal of Zhoukou Normal University

基　　金：河南省自然科学基金(No.14A520042);河南大学科学基金会(No.2012YBZR040)

摘　　要：作为大型科学计算的重要范式,科学工作流通常被部署在高性能计算中心或云平台执行.如何有效执行科学工作流并能满足用户的QOS需求一直是科学计算领域的研究热点.传统的科学工作流调度算法通常以时间或费用为调度目标,而忽略了分布式计算环境中任务执行的安全性问题,即无法保证任务的执行处于一种安全状态或者任务到资源的映射可靠.针对这一问题,考虑任务与资源间的信任关系以及资源的损失效率,提出一种新的信任度量化任务在资源上执行的安全性.在此基础上,鉴于化学反应优化算法的有效性,对化学反应优化算法进行了改进并应用于以执行时间、安全性为目标的多目标工作流调度问题中,提出了一种多目标工作流调度算法MOCRO.基于工作流仿真模拟器Workflowsim的实验结果表明,相对于MOGA、MOPSO算法,MOCRO算法在求解方面具有较好的性能.As an important paradigm for large-scale scientific computing,scientific workflow is usually deployed in high performance computing center or cloud platform implementation.How to effectively implement the scientific workflow and to meet the needs of users of QOS has been a hot research topic in the field of scientific computing.The traditional scientific workflow scheduling algorithm usually in time or cost as the scheduling objective,while ignoring the distributed computing environment to perform tasks in the safety problem,namely it is unable to guarantee execution of tasks in a safe state or task to resource mapping and reliable.In view of this problem,considering the trust relationship between tasks and resources and the failure rate of resources,a new trust is proposed to quantify the security of the task in the implementation of resources.On this basis,in view of the effectiveness of the chemical reaction optimization algorithm,this paper on the chemical reaction optimization algorithm is improved and applied to the execution time and safety for multi objectives of workflow scheduling problem,puts forward a multi-objective workflow scheduling algorithm MOCRO.Experimental results based on the workflowsim of the workflow simulator show that compared with MOGA,MOPSO algorithm,MOCRO algorithm has better performance in solving the problem.

关 键 词：工作流调度 化学反应算法 多目标 PARETO 信任度 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]