Scheduling Algorithms For Dynamical Real-Time Tasks On Multiprocessor Systems

The purpose of this paper is to study the task scheduling problem of dynamical task sets on the system with DVS multiprocessor. A new task can arrive in the system and the system must do the admission test for the task. If the test is satisfied, the task will be accepted and then it can leave after an interval of execution. In this paper, we propose two kinds of admission control algorithms for the situations when the timing information about the departure time of an executing task is known. The Intuitive Admission Control Algorithm (IACA) is proposed to solve the problem of admission control where the departure time of a task is not known until the end of the active interval for its last job. The Positive Admission Control Algorithm (PACA) for the admission control where the departure time of a task is not known until the execution of its last job is finished. Besides, the related properties are presented and proved. A series of experiments were conducted to evaluate the proposed algorithms. From the experimental results, we can observe that the proposed algorithms with the Best Fit heuristic can reject fewer tasks than that with the First Fit and Worst Fit heuristics. Besides, if the departure time of a task can be known early, the information will be helpful to improve the rejected task numbers.