空间环境中的多处理器混合容错调度算法

摘要

为提高空间应用环境中混合实时任务的容错调度效率,提出基于空间环境的多处理器混合容错调度算法.算法在周期任务调度中采用基于分组的“最佳适应”分配策略,以少量增加的计算时间获得更为紧凑的任务调度结果,从而减少调度所需的处理器数目和任务执行时间;在非周期任务调度中,利用基于空闲时间片的方法完成动态调度,使得混合实时任务的容错调度效率有显著提高,能够更好地适应复杂空间环境的任务处理.仿真结果表明算法在改善混合实时任务容错调度性能上具有重要作用.%In order to improve the efficiency of fault-tolerant scheduling of hybrid real-time tasks in circumstance of space application,a hybrid fault-tolerant scheduling algorithm for multiprocessor in space environment is proposed.In this algorithm,a grouping based Best-Fit assignment strategy in the periodic task scheduling is used to obtain more compact scheduling result by use of a little increasing computation time,so that the number of processors and the task execution time required by the algorithm are saved.In addition,a time slice based approach is used to finish the dynamic scheduling of aperiodic tasks,thus improving the efficiency of fault-tolerant scheduling of hybrid real-time tasks and making the algorithm more suitable for the task processing in complicated space environment.It is shown from the simulation that this algorithm has important effect on improving the performance of fault-tolerant scheduling of hybrid real-time tasks.