Abstract:

Universities often introduce a research component with its own credit points into master’s degree programs, but this is generally performed separately from the courses. However, blending learning with other creative activities, directly related to the taught subjects, may stimulate students to study and to share their findings with colleagues. This paper presents an approach adopted within a course of Model Driven Engineering, in the Systems Engineering domain of study. Some of the traditional lecture / application classes were replaced with research activities, having the specific objectives of identifying a new modelling language and of generating a dedicated editor. The purpose is the formal characterization and the graphical representation of systems developed for a given application domain. Throughout the time, we considered domains like: sensor networks, measuring systems, hazard management systems etc. The general objectives, irrespective of the chosen domain, were to:
(i) get students involved in the study of real-life systems,
(ii) force the capability of abstraction, by identifying the recurrent concepts within a set of examples,
(iii) make them defend their case studies in front of their classmates,
(iv) contribute with ideas in the creation of a new modelling language and
(v) collaborate for specifying a formal definition that supports the generation of a new modelling editor.

The paper gives a concrete example about the application of this approach to hazard management systems, inspired from the experience acquired from the research activity of 70 master's students, organized in three working groups. The systems modeled with the newly created language are related to multiple types of hazards. The students studied early warning and alert systems for: fire detection in an urban environment or in a forest, oil spills, floods, tsunami, landslides induced by heavy rainfalls, volcano, tornado, earthquakes, cyclones etc. Due to the large variety of these phenomena, it is difficult to define a common language that is specific to the hazard management domain, yet capable to distinguish the particularities and the differences between the existing systems.

The method adopted is a combination of individual and team work, and the resulted tools are mostly driven by the students’ findings. After the application domain was chosen, there is a lecture that briefly presents the theoretical background. Afterwards, the students have to study a real-life system, chosen from a large set of examples. They use publicly available documentation, like scientific papers, standard specifications, project deliverables etc., and they have to represent a conceptual diagram, containing classes and relationships, with a general modelling language. After that, there are two collaborative sessions; the former consists of presentations of the individual studies in front of the entire class, so each student finds out all the examples studied by his or her colleagues; the latter is dedicated to the analysis of the entire set of conceptual diagrams, with the aim of deducing abstract concepts and relationships that are general enough for characterizing any of those systems. Based on this, students represent the abstract syntax of a new modelling language with Generic Modeling Environment and they generate a correspondent editor, further used by each student for representing the system chosen for the individual case study.

Keywords:

Model Driven Engineering, Research in Education, Hazard Management.