Ground setting properties for an efficient translation of OCL in SMT-based model finding

Abstract

Model Finding is an established method to increase the confidence in the correctness of a UML/OCL model, e. g., by automatically determining valid system states or counterexamples. In the recent past, numerous approaches have been proposed for this purpose. In order to cope with the underlying complexity, approaches based on satisfiability solvers have been found promising. They require a translation of all OCL constraints of the model for a corresponding solver. In this paper, SMT-based model finding is investigated. It is shown that certain OCL operations are causing huge SMT formulations which harm the solving process. However, this is not necessary if a fixed structure of the model can be assumed. Motivated by this, a new concept called ground setting properties is introduced which allows for an efficient translation of OCL into SMT. This concept is illustrated by means of a running example and compared to existing solutions.