一种基于改进蚁群算法的多目标优化云计算任务调度策略

发布时间：2018-03-18 07:10

  本文选题：云计算　切入点：蚁群算法　出处：《微电子学与计算机》2017年11期 　论文类型：期刊论文

【摘要】：提出了一种兼顾任务最短完成时间、成本和负载均衡的改进的集多目标优化的蚁群任务调度算法(time,cost and load balance ant colony optimization,TCL-ACO).首先,针对云计算下任务调度的特点定义任务完成时间成本的约束函数和负载均衡度函数.对于蚁群算法进行初始信息素、启发函数、信息素更新方式进行改进.然后,用改进的蚁群算法求解目标约束函数得到全局最优解.最后在cloudsim下进行仿真,并与Min-Min算法和ACO算法进行仿真对比,实验结果表明本文算法在成本、任务的执行时间和负载均衡方面优于这两种算法.
[Abstract]:In this paper, an improved ant colony task scheduling algorithm, which takes into account the minimum completion time, cost and load balancing of tasks, is proposed, which is cost and load balance ant colony optimization TCL-ACOO. According to the characteristics of task scheduling in cloud computing, the constraint function and load balancing function of task completion time cost are defined. The initial pheromone, heuristic function and pheromone updating method of ant colony algorithm are improved. The improved ant colony algorithm is used to solve the objective constraint function to obtain the global optimal solution. Finally, the simulation is carried out under cloudsim, and compared with the Min-Min algorithm and ACO algorithm. The experimental results show that the proposed algorithm is cost effective. Task execution time and load balancing are better than these two algorithms.
【作者单位】： 重庆邮电大学计算机科学与技术学院;重庆邮电大学图书馆;
【分类号】：TP18;TP3
，

本文编号：1628508