The improved EDF scheduling algorithm for embedded real-time system in the uncertain environment

Most of real-time task scheduling algorithm of embedded real-time systems are based on the worst time for the implementation of scheduling, which will be a waste of resources. Aiming at network system for real-time multimedia application requirements and the current scheduling algorithm lack of real-time performance(ability), this paper analyzes the typical dynamic EDF(Eealiest Deadline First) scheduling algorithm for embedded real-time system, establishes a periodic task set model and presents an improved EDF scheduling algorithm based on the model. Through theoretical analysis and simulation results, in the non-overload case, the algorithm can achieve the EDF scheduling algorithm considerable performance, and mean switching number among tasks is less than the EDF algorithm; in overload case, compared with the EDF algorithm, it has more better scheduling performance, which can effectively guarantee the scheduling success of the high-priority task; finally, this verifys the effectiveness of the improved EDF scheduling algorithm.