An Online Scheduling Algorithm for Reconfigurable Tasks Based on Dynamic Planning Preemptive Threshold

In FCFS and planning scheduling algorithms, because of the selfishness and exclusiveness of the task´s individual scheduling behavior, it is hard to realize a friendly scheduling between two consecutive tasks. This paper gives an online scheduling algorithm based on dynamic planning preemptive threshold. According to the deadline of the new arriving task, a planning preemptive threshold of the planned task queue is established to avoid preempting the tasks whose planning start time exceeds the threshold unnecessarily. In this paper, a concept of urgency of tasks is brought forward. When the planned task´s planning start time is earlier than the threshold, the new arrival task would preemptive the planned tasks by maxima urgency first strategy method, and it optimizes the scheduling efficiency among consecutive tasks. The experiments show that the new algorithm increases the probability of successful schedule, and not reduce the efficiency of scheduling system.