摘要

Recently, real-time system was widely applied to diverse environments. In order to meet the demands of those applications, many scheduling strategies were presented to achieve either maximal benefit or minimum miss deadline ratio. However, very little attention has been devoted to simultaneously achieve the two objectives. This paper proposes a dynamic priority assignment (DPA) strategy by analyzing the remainder value density and urgency of task, in which, two parameters p and q are used to adjust the weights of remainder value density and urgency on task's priority. Based on DPA strategy, the condition that can avoid system thrashing is discussed, and a dynamic real-time task scheduling (DRTS) algorithm is also proposed. Finally, experimental results show that the proposed method can improve the integrated performance of real-time system compared with analogous algorithms.