Navegación

Búsqueda

Búsqueda avanzada

Generating Test Systems in Simulink Models for Testing Product Lines with ASTERYSCO

Simulink models are commonly employed to simulate and test complex systems such as Cyber-Physical Systems (CPSs). These systems are becoming highly configurable, and techniques from the product line engineering context (e.g., feature models) are being acquired by industrial practitioners to model the variability. Having variability in these systems means that there might be several configurations to test. Selecting relevant configurations by considering feature models following combinatorial techniques has been widely investigated by the software engineering community. However, efficiently testing each configuration has attracted little attention, which is not that trivial. One important aspect when testing such systems is automation. This tool paper presents ASTERYSCO, which aims at automatically generating test system instances in Simulink for testing specific configurations of configurable CPSs.

Towards Mutation Testing of Configurable Simulink Models: a Product Line Engineering Perspective

Mutation testing has been found to be an efficient technique in order to assess the quality of a test suite. The use of Simulink models is increasing in both industry and academia to model and simulate complex systems such as Cyber-Physical Systems (CPSs). An advantage of Simulink is its ease to integrate software and control algorithms with complex mathematical models that typically represent continuous dynamic behaviors. In addition to that, the increasing trend of industry in adopting product line engineering methods to efficiently support the variability that their products demand is resulting in configurable Simulink models. Consequently, many configurations can be employed to test the configurable system. Each of these configurations will have a set of mutants, which will be in accordance with the configuration characteristics (i.e., features). However, manually generating and configuring mutants for each of the configurations is a time-consuming and non-systematic process. To deal with this problem, we propose a methodology supported by a tool that automatically generates mutants for configurable Simulink models.