ModiGen introduces a novel workflow that uses large language models to generate Modelica code for complex physical system simulations, reducing the need for specialized expertise.

• Creates benchmark datasets to evaluate LLM performance in Modelica component model generation
• Develops a multi-task workflow for automated creation of simulation code
• Demonstrates how AI can bridge the gap between domain knowledge and modeling implementation
• Provides a path to make advanced engineering simulation more accessible

This research significantly impacts engineering workflows by allowing more engineers to create and optimize physical system models without deep Modelica expertise, potentially accelerating design cycles and innovation in industrial applications.

ModiGen: A Large Language Model-Based Workflow for Multi-Task Modelica Code Generation