Classifying assumptions made during requirements verification of embedded systems

Abstract

We are investigating ways to improve the process of modelling of embedded systems for formal verification. In the modelling process, we make a mathematical model of the system software and its environment (the plant), and we prove that the requirement holds for the model. But we also want to have an argument that increases our confidence that the model represents the system correctly (with respect to the requirement). Therefore, we document some of the modelling decisions in form of a list of the system assumptions made while modelling. Identifying the assumptions and deciding which ones are relevant is a difficult task and it cannot be formalized. To support this process, we give a classification of assumptions. We show our approach on an example. © 2008 Springer-Verlag Berlin Heidelberg.