Simulated restriction of coverage criteria on UML state machines

Model-based testing is a promising quality assurance technique. Automatic test generation from behavioral models is state of the art. Coverage criteria at model level are often used to measure test quality and to steer automatic test generation. While tests are focused on behavior, however, most coverage criteria are focused on the structure of test models. Thus, semantic-preserving model transformations can be used to alter the effect of the applied coverage criteria. In previous work, we presented the notion of simulated satisfaction to enhance the effect of applicable coverage criteria. In this paper, we present model transformations to restrict the effect of coverage criteria. The aim is to show the relative strength of coverage criteria and their dependence on the model structure. The long-term objective is to define lower boundaries for coverage criteria.