Design Structure Matrix Modularization with Large Language Models

TL;DR

This paper extends LLM-based methods to DSM modularization, achieving near-reference quality efficiently and revealing that domain knowledge can hinder performance in complex cases due to semantic misalignment.

Contribution

It introduces a novel LLM-based approach for DSM modularization, analyzing the impact of domain knowledge and proposing the semantic-alignment hypothesis.

Findings

Achieves near-reference quality within 30 iterations

Domain knowledge impairs performance on complex DSMs

Effective input representation and solution pool design are identified

Abstract

Design Structure Matrix (DSM) modularization, the task of partitioning system elements into cohesive modules, is a fundamental combinatorial challenge in engineering design. Traditional methods treat modularization as a pure graph optimization, without access to the engineering context embedded in the system. Building on prior work on LLM-based combinatorial optimization for DSM sequencing, this paper extends the method to modularization across five cases and three backbone LLMs. Our method achieves near-reference quality within 30 iterations without requiring specialized optimization code. Counterintuitively, domain knowledge, beneficial in sequencing, consistently impairs performance on more complex DSMs. We attribute this to semantic misalignment between the LLM's functional priors and the purely structural optimization objective, and propose the semantic-alignment hypothesis as a testable condition governing knowledge effectiveness with LLMs. Ablation studies identify the most effective input representation, objective formulation, and solution pool design for practical deployment. These findings offer practical guidance for deploying LLMs in engineering design optimization.