Agentic Llm Architecture For Real-Time Facility Layout Optimization With An Organized Llmops Framework
Keywords:
Agentic LLM, Digital twin, Facility layout planning, LLMOps, Real-time optimizationAbstract
Facility layout planning is a combinatorial industrial engineering problem in which material handling cost, travel distance, space utilization, safety, and production flow must be considered simultaneously. Conventional optimization methods can obtain high-quality solutions, but they generally require manual formulation, algorithm selection, parameter tuning, and repeated interpretation of operational changes. This study proposes an agentic large language model (LLM) architecture coupled with an organized LLM Operations (LLMOps) framework for real-time facility layout optimization. The architecture separates planning, optimization, verification, simulation, and critique into specialized agents while using external optimization tools rather than asking the LLM to perform numerical optimization directly. A synthetic benchmark containing 40 layout scenarios with demand and flow perturbations was used to compare four approaches: manual systematic-layout-planning (SLP) heuristic, genetic algorithm (GA), single tool-augmented LLM, and the proposed agentic LLM architecture. The proposed method achieved a mean material-handling-cost index of 88.39, compared with 128.54 for manual SLP, 110.54 for GA, and 103.83 for the single-LLM baseline. This corresponds to a 31.2% reduction relative to manual SLP and a 20.0% reduction relative to GA. Mean logistics distance decreased by 26.8% relative to the manual baseline. Across the benchmark, the proposed workflow produced a feasible layout in 95% of scenarios without manual repair. A paired comparison against the manual baseline was statistically significant (t=46.86, p<0.001). The findings indicate that the main contribution of agentic LLMs is not replacing established optimization algorithms, but coordinating data interpretation, tool invocation, constraint verification, simulation feedback, and reproducible model operations within one auditable workflow.
References
Bajoria, S., Ranjan, S., Adhitya, M., Suri, V., Hassija, V., Chamola, V., & Hussain, A. (2026). From Language Models to Agentic AI: A Survey of Autonomous, Action-Enabled, and Collaborative LLM Agents. Cognitive Computation, 18, 103. https://doi.org/10.1007/s12559-026-10619-1
Choi, H., Yu, S., Lee, D. H., Noh, S. D., Ji, S., Kim, H., Yoon, H., Kwon, M., & Han, J. (2024). Optimization of the Factory Layout and Production Flow Using Production-Simulation-Based Reinforcement Learning. Machines, 12(6), 390. https://doi.org/10.3390/machines12060390
Du, S., Zhao, J., Shi, J., Xie, Z., Jiang, X., Bai, Y., & He, L. (2026). A Survey on the Optimization of Large Language Model-based Agents. ACM Computing Surveys, 58(9), 223. https://doi.org/10.1145/3789261
Feldt, J., Kourouklis, T., Kontny, H., & Wagenitz, A. (2020). Digital twin: Revealing potentials of real-time autonomous decisions at a manufacturing company. Procedia CIRP, 88, 185–190. https://doi.org/10.1016/j.procir.2020.05.033
Guo, H., Zhu, Y., Zhang, Y., Ren, Y., Chen, M., & Zhang, R. (2021). A digital twin-based layout optimization method for discrete manufacturing workshop. The International Journal of Advanced Manufacturing Technology, 112, 1307–1318. https://doi.org/10.1007/s00170-020-06568-0
Guo, X., Huang, Y., Zhang, S., & Ma, G. (2026). Three-dimensional facility layout optimization using large language model-driven A-star and non-dominated sorting genetic Algorithm-II. Expert Systems with Applications, 315, 131745. https://doi.org/10.1016/j.eswa.2026.131745
Isah, H., Abughofa, T., Mahfuz, S., Ajerla, D., Zulkernine, F., & Khan, S. (2019). A Survey of Distributed Data Stream Processing Frameworks. IEEE Access, 7, 154300–154316. https://doi.org/10.1109/ACCESS.2019.2946884
Jazi, A. E., & Ameri, F. (2026). Machine learning, digital twin, and AR/VR in manufacturing facility layout planning: a review. International Journal on Interactive Design and Manufacturing. https://doi.org/10.1007/s12008-026-02629-y
Kroemer, O., Niekum, S., & Konidaris, G. (2023). Deep reinforcement learning for layout planning – An MDP-based approach for the facility layout problem. Manufacturing Letters, 38, 40–43. https://doi.org/10.1016/j.mfglet.2023.09.007
Lanzini, M., Ferretti, I., & Zanoni, S. (2024). Towards the Implementation and Integration of a Digital Twin in a Discrete Manufacturing Context. Processes, 12(11), 2384. https://doi.org/10.3390/pr12112384
Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W.-t., Rocktäschel, T., & Kiela, D. (2020). Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks. Advances in Neural Information Processing Systems, 33. https://doi.org/10.48550/arXiv.2005.11401
Li, H., Zhang, H.-Y., Liao, Y., Tian, J.-M., & Kuang, M.-S. (2025). Facility layout optimisation of an intelligent manufacturing unit based on queueing network and deep reinforcement learning. International Journal of Production Research, 64(4), 1295–1318. https://doi.org/10.1080/00207543.2025.2568181
Ma, D., Lang, S., Kute, S., Reider, R., & Müller, M. (2026). LLMOps for End-to-End Automation in Facility Layout Planning: Methodology and Application. Procedia Computer Science, 277, 1538–1547. https://doi.org/10.1016/j.procs.2026.02.192
Plaat, A., van Duijn, M., van Stein, N., Preuss, M., van der Putten, P., & Batenburg, K. J. (2025). Agentic Large Language Models, a Survey. Journal of Artificial Intelligence Research, 82. https://doi.org/10.1613/JAIR.1.18675
Santos, R., Basto, J. P. T. V., Alcalá, S. G. S., Frazzon, E. M., & Azevedo, A. (2019). Industrial IoT integrated with simulation: A digital twin approach to support real-time decision making. In EU IEOM 2019: Proceedings of the 3rd European International Conference on Industrial Engineering and Operations Management (pp. 816–828). https://doi.org/10.46254/EU03.20190225
Schick, T., Dwivedi-Yu, J., Dessì, R., Raileanu, R., Lomeli, M., Zettlemoyer, L., Cancedda, N., & Scialom, T. (2023). Toolformer: Language Models Can Teach Themselves to Use Tools. Advances in Neural Information Processing Systems, 36. https://doi.org/10.48550/arXiv.2302.04761
Wang, L., Ma, C., Feng, X., Zhang, Z., Yang, H., Zhang, J., Chen, Z., Tang, J., Chen, X., Lin, Y., Zhao, W. X., Wei, Z., & Wen, J.-R. (2024). A survey on large language model based autonomous agents. Frontiers of Computer Science, 18(6), 186345. https://doi.org/10.1007/s11704-024-40231-1
Yao, S., Zhao, J., Yu, D., Du, N., Shafran, I., Narasimhan, K., & Cao, Y. (2023). ReAct: Synergizing Reasoning and Acting in Language Models. International Conference on Learning Representations (ICLR 2023). https://doi.org/10.48550/arXiv.2210.03629
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Unified Systems Approach for Human centered Operations (USAHO)

This work is licensed under a Creative Commons Attribution 4.0 International License.

