Automated Design Structure Matrix Generation with LLMs and GraphRAG

A recent study published on arXiv examines the application of Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) techniques, specifically GraphRAG, for the automated generation of Design Structure Matrices (DSMs). DSMs are tools used in systems engineering to visualize and manage dependencies between components of a complex system.

The research focuses on two specific use cases: the design of a power screwdriver and a CubeSat. The goal is to evaluate the effectiveness of these methods in identifying relationships between predefined components and, more importantly, in discovering new components and their interconnections.

Preliminary results, while highlighting design and computational challenges, suggest that automating DSM generation via LLMs and GraphRAG is a promising area. The code used in the study is publicly available for reproducibility and to encourage further development by domain experts.

For those evaluating on-premise deployments, there are trade-offs to consider. AI-RADAR offers analytical frameworks on /llm-onpremise to evaluate these options.