针对网格环境下具有约束关系的任务调度问题,基于有向无环图DAG(directed acyclic graph)设计了调度模型;提出了一种改进的退火进化算法,对任务的执行次序和资源的具体分配分离编码,给出适应度函数计算方法和算法步骤。最后将算法和传统的遗传算法比较,实验结果显示该算法能获得更好的调度结果。
According to the constraint relation task scheduling problem,this paper designed the scheduling model based on directed acyclic graph(DAG).And proposed an improved annealing evolutionary algorithm,separated coding the task execution order and resource allocation,gave fitness function calculation method and algorithm procedures.Compared with the traditional genetic algorithm,experimental results show that this algorithm can obtain better scheduling results.