earticle

영어

Focusing on the problem that users would remove or delete the task at any time in cloud computing environment, a new scheduling algorithm for adapting to dynamic changes of user task is proposed. At first, the algorithm uses directed acyclic graph (DAG) to describe the association relation of cloud computing tasks. When a task needs to be removed, the algorithm utilizes a cost function to determine whether to cancel this task, and then updates the DAG according to its dependence on removed task. Finally, it uses a heuristic algorithm to perform scheduling. Experimental results show that this algorithm can better avoid scheduling execution of revoked tasks and improve the executive efficiency of cloud computing resource. Its performance is better than Min-Min and Max-Min in executive time span.