Transition-based testability analysis for reactive systems

Reactive systems are often designed as two parts: computation part and control part. The computation part is modeled by operator diagrams, while the control part is modeled by transition-based models. In this paper, we concentrate on analyzing testability of control part by basing upon transition-based models. We first transform transition-based models into Markov chains by augmenting probability information. Then, testability measures are proposed from Markov chains as an estimate of testing effort for reaching state coverage and path coverage. The approach is applied to a case study and the obtained measures are compared to the testing effort required by a test generation tool. The results shows some interesting perspectives.