Optimal scheduling algorithm for real-time tasks in digital control systems based on GA

Abstract

On the basis of analysis on the characteristic of the tasks in control system, the task model and subtask model are given. A scheduling algorithm for control system named half-preemptive EDF algorithm is proposed. The algorithm can guarantee the utility of processor and reduce the times of tasks that are preempted by other tasks. Because the performance of control system is influenced by sampling period and control delay and they affect each other, genetic algorithm is adopted to optimize the performance of control system in order to acquire optimal sampling period. Based on floating-point coding, elitist species and rank based selection are combined. The simulation results show that the algorithm presented in this paper can guarantee schedulability of tasks and improve the performance of control system.