遗传算法在网格任务调度的应用研究

摘要

研究了网格任务调度问题.针对传统任务调度算法在网格环境下存在不能很好地平衡节点负载和满足用户服务质量需求等缺点,导致网格系统负载极不均衡,调度效果低.为了提高网格任务调度的效果,提出一种基于遗传算法的网格任务调度方法.将网格任务编码成种群中的个体,网络任务目标作为遗传算法的适应度函数,通过遗传算法的强全局搜索及交叉、变异操作,获得最优的任务调度方案.仿真结果表明,采用遗传算法进行网格任务调度可以减少系统总执行时间和任务完成时间,提高了资源调度效率,使网格系统负载均衡度更好,在网格任务调度具有广泛的应用前景.%Research grid task scheduling problem. Traditional task scheduling algorithms are unsuitable for the grid environment for unbalanced load and unsatisfied customer service quality requirements, with the result that network system load is unbalanced and scheduling rate is extremely low. In order to improve the effect of grid task scheduling , a grid scheduling method is proposed based on genetic algorithm. This method encodes population of individuals for grid task, takes the network task goal as the fitness function of the genetic algorithm, and the strong global search ability of genetic algorithm is used and the operations of crossover and mutation are carried out to obtain the optimal task scheduling scheme. Simulation results show that, using genetic algorithms for grid task scheduling can reduce the system total execution time and task completion time, improve resource scheduling efficiency, and make a grid system load balance degree better. It has wide application prospects in the grid scheduling.