The transition from traditional, mono-disciplinary products to complex, multidisciplinary smart products has fundamentally reshaped engineering practices. Model-Based Systems Engineering (MBSE) and Artificial Intelligence (AI), particularly in combination, have emerged as key enablers for managing this complexity by facilitating the automated, generative creation of system models. Despite their potential, the adoption of these approaches is often impeded by a lack of user expertise, technical complexity, and organizational resistance. At the same time, context-specific prompting itself can have significant influence on the generated code quality. This paper presents a comprehensive, multi-layered framework for generative systems engineering that systematically integrates context-specific supporting prompt engineering, Retrieval Augmented Generation, and Large Language Models to support the automate creation of high-quality, context-aware system models. The framework addresses requirements of various engineering roles by orchestrating user intent, engineering methodologies, tool interoperability, and domain-specific knowledge sources across five interrelated layers. By enabling structured natural language prompting, the framework lowers technical barriers and reduces resistance to MBSE adoption, making AI-driven modeling more accessible to users with varying backgrounds. The framework’s effectiveness is demonstrated through its instantiation in an e-bike case study, showcasing its ability to generate SysML v2-compliant requirements models directly within industry-standard tools.
Mollahassani et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: