Combining Model-Based Systems Engineering and Knowledge-Centric Systems Engineering to Design Reliable Systems in Practice

The use and importance of complex software-intensive systems are growing. As they are used in a wider range of situations in which dependability must be ensured, the reliability of the systems and of their components needs to be addressed throughout their lifecycle, including at early development stages. In addition, the means used to deal with reliability need to be linked to and integrated into the overall systems engineering practices and processes. Within this context, we present an approach to design reliable systems in practice in the scope of model-based systems engineering (MBSE) and knowledge-centric systems engineering (KCSE), two systems engineering perspectives whose adoption is increasing. While MBSE relies on explicit system models, KCSE places artificial intelligence at its core to capture, formalise, and reason over system knowledge. Both perspectives are combined to model systems and analyse whether their design addresses the expected system reliability properties, leveraging knowledge representation, natural language processing, and inference mechanisms. The approach links the processes and tools of Arcadia/Capella for MBSE and of SES Engineering Studio for KCSE. A joint application process has been defined for system modelling, ontology development, structured textual requirements specification, traceability management, and model quality analysis, all of which are targeted at system reliability. For validation, the approach has been applied on eight systems that cover five different application domains, considering tens of diagrams, of knowledge elements, of reliability properties, and of analysis possibilities. Based on the validation results, we argue that the approach is a feasible means to design reliable systems. The approach is also the first one that effectively combines MBSE with Arcadia/Capella and KCSE with SES to design reliable systems in practice.