TOWARDS AUTOMATED ANALYSIS OF EXECUTABLE MODELS

Detta är en Master-uppsats från Mälardalens högskola/Akademin för innovation, design och teknik

Författare: Jean Malm; Jonas Skoog; [2017]

Nyckelord: ;

Sammanfattning: By utilising graphical modelling languages, software developers can design large systems while keeping the architecture simple and understandable. Another benefit of using modelling languages is the possibility to automatically generate code, saving both time and reducing human errors in code. In addition, through the usage of well defined and executable models, the system can be tested and analysed as early as during the design phase. Doing so, potential errors in the design can be found and corrected during the design phase, saving time and other resources. In this thesis we present a method for analysing executable model representations of systems through three steps. First the model is transformed to a suitable format for analysis. The result of the transformation is then analysed using analysis tools. Lastly, the result of the analysis is mapped back to the source model. The developed method was assessed through a proof-of-concept using a transformation from the action language for fUML, an executable modelling language. The target of the transformation was ALF, an intermediate language designed for analysis in the SWEET tool. The analysis of the source was performed in regards to properties of the system, such as program flow and value bounds. Through the proof-of-concept, the proposed method was shown to be a viable option for analysing fUML models.

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)