A New Adaptive Performance Feedback Control Scheduling Model Oriented to the Embedded Soft Real-Time Systems

The running characteristics of tasks are uncertain in an unpredictable environment. So the workload of a soft real-time system is time varying, too. An adaptive performance control model which combines the PID and fuzzy feedback technologies is presented. When the system performance deviation is lower than the threshold value, the adaptive feedback control technology based PID is adopted , or the adaptive fuzzy feedback technology is used. Meanwhile,we introduce the fuzzy control decision table instead of fuzzy reference computation to reduce CPU computation overhead in fuzzy feedback control. The result of simulation test shows that the fuzzy-PID feedback control model needs much less time to reach steady state than that of the adaptive PID performance control model when the actual system workload dramatically and abruptly varies.