A comparative study of languages for model-based systems-of-systems engineering (MBSSE)

Abstract

Teaching model-based systems engineering has become more challenging as research and development has shifted from systems to include systems-of-systems (SoSs). SoSs involve primarily combination of hardware and software, as well as humans and organizations. The need to teach model-based system-of-systems engineering has thus raised the need to determine what the most suitable language for modeling and teaching SoSs is. We developed a new course format, in which groups jointly reverse-engineer and model Web-based information systems in two different modeling languages, UML and OPM, and then individually assess their peers' projects. The goal of this study was to compare UML with OPM as two candidates. We developed and evaluated an online peer assessment tool which students used. About 130 undergraduate students in groups of six divided into teams of three modeled 23 systems in both UML and OPM. They then evaluated, compared, and ranked the clarity & understandability and the completeness of the four models of two systems that their peers had constructed. Findings demonstrate that neither the order of the models assessment nor the assessor gender affected the grading, indicating assessment reliability. We found significant differences in model clarity and understandability in favor of OPM and no differences in completeness between OPM and UML models. These findings validate our approach of teaching both conceptual modeling languages and using peer assessment in large-scale project-based undergraduate information systems engineering courses.