Scalable Scenario Specifications to Synthesize Component-centric Behaviour Models

Abstract

Several scenario description languages and associated behaviour synthesis processes have been developed. The goal of these is to synthesize behaviour models from system requirement specifications, in order to enable the early identification of weak design spots or code generation. To date, however, most of the scenario languages are poorly scaled with regards to system sizes. This is because scalability requires more expressive constructs that can help when writing a scenario specification in a concise and compact manner, thereby resulting in a reduced number of scenarios. Furthermore, due to a lack of expressiveness in scenario languages, synthesis algorithms may need to rely on global behaviour models to determine inter-scenario dependencies. The global model is an additional factor that limits the approaches' scalability. The reason is that the construction of a global model becomes harder as the system specification increases. To tackle these issues, within this article is proposed an expressive scenario description language that provides a concise and compact approach to scenario description, and defines inter-scenario dependencies semantically. A new algorithm that can address the additional constructs of the language was defined, in order to synthesize component-centric behaviour models. The applicability of this work has been demonstrated through both an illustrative example and a real-world case study. The evaluation indicates that the proposed scenario description language is more scalable than existing languages.