摘要
针对目前高性能计算任务调度策略利用率低、负载不均衡等问题,设计一种基于遗传-蚁群算法的高性能计算任务调度算法(GA-ACO)。GA-ACO分为两个阶段,第一阶段通过遗传算法缩小空间快速搜索到优秀解,紧接着将其转化为蚁群算法的初始信息素;第二阶段提出一种基于蚁群信息素的全局更新策略对收敛速度做出优化。实验分析表明,与蚁群算法和遗传算法相比,该算法缩短了任务完成时间,降低了节点负载率。
Aimed at the problems of low utilization rate and unbalanced load of current high performance computing task scheduling strategies,a high-performance computational task scheduling algorithm based on genetic ant colony optimization(GA-ACO)is designed.GA-ACO was divided into two stages.In the first stage,the genetic algorithm was used to narrow the space and quickly find the excellent solution,and then it was transformed into the initial pheromone of ant colony algorithm.In the second stage,a global update strategy based on ant colony pheromone was proposed to optimize the convergence speed.Experimental analysis shows that compared with ant colony algorithm and genetic algorithm,this algorithm shortens the task completion time and reduces the node load rate.
作者
田智慧
张帅永
高需
Tian Zhihui;Zhang Shuaiyong;Gao Xu(School of Earth Science and Technology,Zhengzhou University,Zhengzhou 450052,Henan,China;School of Information Engineering,Zhengzhou University,Zhengzhou 450001,Henan,China;Zhengzhou National Supercomputing Center,Zhengzhou 450001,Henan,China)
出处
《计算机应用与软件》
北大核心
2024年第3期253-257,共5页
Computer Applications and Software
基金
国家重点研发计划项目(2018YFB0505004-03)
郑州大学2018年科研启动基金项目(32210919)。
关键词
高性能计算
任务调度
遗传算法
蚁群算法
信息素
High performance computing
Task scheduling
Genetic algorithm
Ant colony algorithm
Pheromone
