针对云计算中现有任务调度模型为追求最短完成时间,而没有从物理资源的负载均衡角度考虑,提出了基于任务预先分类的调度策略(PCSS)。该策略通过综合考虑任务的最少执行时间和截止时间来设置优先级,根据优先级别将任务批量分配给合适的物理机,实现了任务的最短完成时间和物理机的负载均衡。最后通过Cloudsim仿真实验分析和比较,该策略能很好地执行用户任务并体现出良好的负载均衡。 In view of the existing task scheduling model in cloud computing, in order to pursue the shortest completion time without considering the load balance of physical resources, a scheduling policy (PCSS) based on task pre-classification is proposed. The strategy sets the priority by considering the minimum execution time and deadline of tasks, and allocates the tasks to the appropriate physical machines according to the priorities, achieving the task minimum completion time and physical machine load balancing. Finally, through the analysis and comparison of Cloudsim simulation experiments, the strategy can perform user tasks well and reflect good load balancing.