Fuzzy rule-based approach to real-time scheduling

In real-time systems, tasks have to be performed not only correctly, but in a timely fashion. Task scheduling is essential for designing a real-time system, because the scheduling algorithm ensures if tasks meet their deadlines. However, the inherit nature of uncertainty in dynamic real-time systems increases the problems inherent in scheduling. To alleviate these problems, we proposed a fuzzy scheduling approach in which the real-time scheduling problem is treated as a multi-criteria optimization problem, and a set of fuzzy rules is utilized to derive a feasible schedule. A simulation is also conducted to evaluate the performance of the proposed approach. The result of the simulation shows that the proposed fuzzy scheduler performs very closed to the optimal minimum laxity first (MLF) in terms of task loss, and performs significantly better than MLF in choosing important tasks to execute.