Abstract The digital transformation of systems engineering requires converting legacy documentation into standardized models, but manual conversion remains time‐intensive and error‐prone. The recent development of Large Language Models (LLMs) offers an exciting but unexplored solution. This paper conducts a mathematical analysis to show that LLMs can scale in a trivially parallelizable manner with O (n²) time and cost complexity, proving that in general, LLM‐based approaches can theoretically scale up to large systems. Then, the paper introduces a novel automated end to end pipeline that transforms unstructured documentation into SysML Block Definition Diagrams using LLMs, a Groovy script to import the diagrams into Cameo 2022x, and graph theory‐based invariants to improve the output. The system employs proposition‐based retrieval‐augmented generation with smart transitive reduction for relationship refinement, achieving F1 scores of 0. 95 for element identification and 0. 85 for relationship extraction on test documents. This work represents a significant step toward automating the transition from document‐based to model‐based systems engineering, potentially reducing the time and effort required for digital transformation of systems engineering processes.
Johnson et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: