A Novel Intelligent Algorithm for Fault-Tolerant Task Scheduling in Real-Time Multiprocessor Systems
Golbarg Zarinzad, Amir Masoud Rahmani, Nikta Dayhim · 2008
The scheduling problem for real-time tasks on multiprocessor is one of NP-hard problems. In fault-tolerant real-time systems, tasks have deadlines to be met in spite of the presence of fault. Many attempts such as classical algorithms and intelligent methods have been made to solve this problem. Primary-backup (PB) scheme is one of the most important classical algorithms that have been employed for fault-tolerant scheduling for real-time tasks, wherein each task has two versions and the versions must be scheduled on two different processors. In this paper a novel scheduling algorithm is proposed based on genetic algorithm (GA) which uses PB for tolerating faults since all tasks employ the processors equally as much as possible.