摘要
针对云计算处理节点的任务调度问题,提出了一种基于改进离散粒子群算法的云计算任务调度方法;首先,定义了云计算任务调度数学模型,在此基础上对离散粒子群算法进行改进,采用自然数编码来表示任务调度方案对应的粒子位置,提出了一种自适应的惯性权重因子调整方法,并给出了子种群和主种群进行协同寻优的粒子群任务调度算法;仿真实验表明:文中方法获得最优解的次数远大于其他方法,在迭代次数为22次时就获得全局最优解192.34,同时具有良好的收敛特性。
Aiming at the problem of task scheduling of processing node, a task scheduling method was introduced based on improved par- ticle swarm algorism. Firstly, the mathematical model of cloud computing was defined, and then the discrete particle algorism is improved, the natural number coding was used to represent the particle position, and a adaptive inertial weight factor was adjusted, finally, the algorism based on main population and the sub--population cooperation was given. The simulation result shows the counts of obtaining the best solu tion in our method is more than the other methods. When the iteration time is 22, the optimal solution is 192. 34 and have the good property such as convergence.
出处
《计算机测量与控制》
CSCD
北大核心
2012年第11期3070-3072,共3页
Computer Measurement &Control
关键词
任务调度
粒子群
云计算
优化
task scheduling
particle swarm
cloud computing
optimizing
