Hard real-time scheduling for large scale process control applications

Abstract

Real-time systems interact with external environment and hence often have demanding performance requirements. Designing and developing complex real-time systems is a challenging task. Model based programming is a relatively new approach towards integrated system development that simplifies this task. Hard real-time requirements pose additional problem that the timing constraints imposed must be met, or else some catastrophic failure occurs. The authors demonstrate that hard real-time requirements can be satisfied by a suitable scheduling scheme in model based programming environment. The scheduling scheme consists of a scheduling algorithm and schedulability tests. Both the scheduling algorithm and the schedulability tests can be performed in polynomial time, hence can be used in dynamic environment where on-line restructuring of the computation model may be necessary.<>