Constraint Programming in Embedded Systems Design: Considered Helpful

Abstract Embedded systems are built for specific purposes and are optimized to meet different kind of constraints, such as performance, timing, power and cost. The design process therefore involves different optimization activities. In this paper, we discuss the use of constraint programming (CP) technology for these optimization problems. The main advantages and disadvantages of applying CP to embedded system design problems are discussed on two examples, scheduling and mapping. Based on these examples modelling capabilities of CP and basic solving methods are discussed. We have identified CP modelling capability as an important factor for problem formalization and their uniform representation. We have also, using several experiments, show efficiency of the models and solving process. Finally, we have also pointed out difficulties with CP technology that are mostly related to search methods that, for more realistic problems, must be carefully selected or even new methods must be developed.