LLMOps for End-to-End Automation in Facility Layout Planning: Methodology and Application

Efficient facility layout planning and rapid solution development are crucial for enhancing internal logistics, productivity, and supply chain resilience. Despite extensive research using advanced heuristics or simulation tools, current approaches tend to emphasize single-task performance rather than workflow integration, limiting their industrial applicability. Large Language Model Operations (LLMOps) aims to orchestrate comprehensive end-to-end workflows among models and tools, making it suitable for addressing the interdependent complexities in planning tasks. This paper presents a novel integrated method for facility layout planning that combines the data processing capabilities of Large Language Models (LLMs) with external function calling. The approach creates multiple specialized agents responsible for interpreting technical drawings, implementing optimization algorithms, and summarizing solutions. These agents are orchestrated through an LLMOps platform into a unified workflow. We detail the design and function of each node in the proposed workflow and validate the approach with a conceptual example based on multi-row facility layout problem. This approach illustrates how agents, LLMs, and external tools can be orchestrated to unify fragmented tasks across teams, paving the way for more robust and reusable industrial automation frameworks.