Empowering Model‐Based Systems Engineering Through Metamodeling

Abstract

AbstractA critical enabler for Model‐Based Systems Engineering (MBSE) and Digital Engineering (DE) is the generation of coherent and consistent views of a system‐of‐interest based on information within a system model. In practice, system model development is facilitated through domain‐specific profiles, style guides, reference models, and low‐fidelity meta‐models to create coherent and consistent system information. Each of these approaches are useful but are insufficient for robust and automated verification of system models to an ideal. Furthermore, the expression of domain‐specific concepts and semantics relies on the proliferation of non‐standard, domain‐specific profiles as standard system modeling languages like the Systems Modeling Language (SysML) are general purpose. This paper proposes a novel approach to creating precise, machine‐interpretable metamodels implemented as a lightweight Unified Modeling Language (UML) profile. The profile includes numerous features that allow model architects to quickly specify context and domain‐specific modeling constructs without creating non‐standard stereotypes to apply domain‐specific meaning and usage rules. Three kinds of constraints can be inferred based on the relationships between meta‐model elements: type, multiplicity, and default value. Applications of well‐formed metamodels include a shared understanding of the intended model format and structure, as well as the one‐time programmatic generation of an encompassing suite of validation rules to evaluate a system model against the inferred constraints, thus ensuring consistency. Additional applications include programmatic generation of model analysis metrics, system models from metamodels, metamodels from reference models and element finding queries, and the ability to update a system model based upon the updated metamodel automatically. Use of the approach results in reduced time in system model development and analysis and ensures coherency and consistency of information thus increasing stakeholder use and confidence in the system model.