This study looks at how artificial intelligence, specifically GPT-4, can help speed up the creation of system dynamics simulations in systems engineering. We explored how GPT-4 can assist humans in building and improving these models by performing tasks like creating, expanding, and stabilizing simulations, finding errors, suggesting new ideas, and converting models into Python code. Our findings show that while GPT-4 isn't perfect, it can make the modeling process faster, reduce mistakes, and help people learn more quickly. However, human expertise is still crucial for evaluating and testing these models. We suggest that AI and human modelers should work together, using Large Language Models like GPT-4, to augment human abilities and advance system dynamics and systems engineering.

C. du Plooy & R. Oosthuizen

View details: https://sajie.journals.ac.za/pub/article/view/2944