| 2026 | |
| [1] | , "A Modular Multi-Agent LLM Architecture for Text-to-Diagram Generation and User-Guided Refinement", In e-Informatica Software Engineering Journal, vol. 20, no. 1, pp. 260109, 2026.
DOI: 10.37190/e-Inf260109. Download article (PDF)Get article BibTeX file |
Authors
Hartwig Grabowski
Abstract
Context: Recent advances in LLM-based diagram generation increasingly rely on coordinated agent systems rather than single-model prompts.
Objective: This work highlights how modular multi-agent architectures improve reliability, semantic grounding, and iterative refinement in text-to-diagram workflows.
Method: We analyze a pipeline composed of specialized agents for interpretation, synthesis, validation, and correction, each contributing a bounded and inspectable transformation.
Results: The agent system provides deterministic validation, structured reasoning, and controlled refinement loops that outperform monolithic LLM generation.
Conclusions: Multi-agent LLM pipelines represent a robust foundation for precise, verifiable diagram generation and serve as a reproducible alternative to single-pass text-to-diagram models.
Keywords
diagram generation, multi-agent LLM pipeline, engineering informatics, automated modeling, computational design
References
1. E. Brito, R. Póvoas, and P. Providência, “Drawing in the engineering design process,” in Structures and Architecture . CRC Press, 2016, pp. 871–878.
2. A. Taraszkiewicz, “Freehand drawing versus digital design tools in architectural education,” Global Journal of Engineering Education , Vol. 23, No. 2, 2021, pp. 107–112.
3. K.D.D. Willis, Y. Li, Y. Sun, and T. Funkhouser, “Engineering sketch generation for computer-aided design,” 2021, arXiv:2104.09621.
4. R. Agrawal, K. Jagtap, and R. Kantipudi, “An overview of hand-drawn diagram recognition methods and applications,” 2023, researchGate Preprint.
5. L. Donnici, “Rethinking sketching: Integrating hand drawings, digital tools, and generative ai,” Infrastructures , Vol. 9, No. 5, 2025, p. 119.
6. Autodesk Inc., “Autocad software,” 2024. [Online]. https://www.autodesk.com/products/autocad/overview
7. Trimble Inc., “Sketchup software,” 2024. [Online]. https://www.sketchup.com
8. F. Luleci, F.N. Catbas, and O. Avci, “Generative adversarial networks for acceleration data augmentation in structural monitoring,” Journal of Civil Structural Health Monitoring , Vol. 13, No. 1, 2023, pp. 181–198.
9. G.C. Marano, M.M. Rosso, A. Aloisio, and G. Cirrincione, “Review of gans in earthquake engineering,” Bulletin of Earthquake Engineering , Vol. 22, No. 7, 2024, pp. 3511–3562.
10. Z. Rastin, G.G. Amiri, and E. Darvishan, “Gan for damage identification in civil structures,” Shock and Vibration , 2021, p. 3987835.
11. Y. Feng, Y. Fei, Y. Lin, W. Liao, and X. Lu, “Intelligent generative design using rule-embedded gan,” Journal of Structural Engineering , Vol. 149, No. 11, 2023, p. 04023161.
12. J. White, Q. Fu, S. Hays, M. Sandborn, C. Olea et al., “A prompt pattern catalog to enhance prompt engineering,” 2023, arXiv:2302.11382.
13. R.J.A. Buhr, G.M. Karam, C.J. Hayes, and C.M. Woodside, “Software cad: A revolutionary approach,” IEEE Transactions on Software Engineering , Vol. 15, No. 3, 1989, pp. 235–249.
14. J. Rumbaugh, M. Blaha, W. Premerlani, F. Eddy, and W. Lorensen, Object-Oriented Modeling and Design . Prentice Hall, 1991.
15. G. Booch, Object-Oriented Analysis and Design with Applications , 2nd ed. Benjamin/Cummings, 1994.
16. Object Management Group, “Unified modeling language (uml) specification, version 2.5.1,” 2017. [Online]. https://www.omg.org/spec/UML/2.5.1
17. A. Schürr, A. Winter, and A. Zündorf, “The progres approach: Language and environment,” in Handbook of Graph Grammars and Computing by Graph Transformation . World Scientific, 1999, pp. 487–550.
18. R. France and B. Rumpe, “Model-driven development: A research roadmap,” in Future of Software Engineering (FOSE 2007) , 2007, pp. 37–54.
19. M. Mernik, J. Heering, and A.M. Sloane, “When and how to develop domain-specific languages,” ACM Computing Surveys , Vol. 37, No. 4, 2005, pp. 316–344.
20. T. Fischer, J. Niere, L. Torunski, and A. Zündorf, “Story diagrams: A new graph transformation language based on uml and java,” in Theory and Application of Graph Transformation . Springer, 2000, pp. 296–309.
21. S. Ghosh, P. Mukherjee, B. Chakraborty, and R. Bashar, “Automated generation of er diagram from a given text,” in International Conference on Machine Learning and Data Engineering (iCMLDE) , 2018, pp. 91–96.
22. N. Rane, S. Choudhary, and J. Rane, “Integrating bim with generative ai,” Buildings , Vol. 14, No. 1, 2023, p. 220.
23. L. Giray, “Prompt engineering with chatgpt: A guide for academic writers,” Annals of Biomedical Engineering , Vol. 51, No. 12, 2023, pp. 2629–2633.
24. J. Wei, X. Wang, D. Schuurmans, M. Bosma, F. Xia et al., “Chain-of-thought prompting elicits reasoning in large language models,” in Advances in Neural Information Processing Systems , Vol. 35, 2022, pp. 24 824–24 837.
25. S. Yao, D. Yu, J. Zhao, I. Shafran, T. Griffiths et al., “Tree-of-thoughts: Deliberate problem solving with large language models,” in Advances in Neural Information Processing Systems , Vol. 36, 2023, pp. 11 809–11 822.
26. X. Zhang, L. Iturburu, J.N. Villamizar, X. Liu, M. Salmeron et al., “Llm agent for structural drawing generation using react prompting and rag,” 2025, arXiv:2507.19771.