Towards a GPT-Based Lean Manufacturing Consultant for Manufacturing Optimization

Published: 28 May 2024| Version 1 | DOI: 10.17632/fwpwmgd92s.1
Christian Magnus


This repository contains the dataset used and generated during the research for the article with the title "Towards a GPT-Based Lean Manufacturing Consultant for Manufacturing Optimization". Use and generation of the data is decribed in the corresponding article.


Steps to reproduce

The latest available version of ChatGPT 4 with DALL·E, browsing and analysis as of the latest release note being from Feb 13, 2024, was used. As in this release memory was introduced to a limited number of users, this was checked in the user settings. Neither memory, nor temporary chat functions were displayed. To always start new chats without having influencing factors from the previous chat histories, all previous chats were deleted and ChatGPT was told to forget all information from previous chats with the user, prior to starting the actual conversation. To reduce bias, the following rules for the conversations were followed: Rule 1: If it is offered to optimize the manufacturing process, ask for support in the areas that ChatGPT suggests. If it is suggested to use Lean Manufacturing methods, ask for support in using these by asking “how to start”. Rule 2: Follow the instructions step by step as long as correct outputs / answers are generated or until all results are generated and it would be clearly up to the user to collect further information or get into action. Rule 3: If no specific tool/method is given for the current step, ask for these. If several tools/methods are given, always ask for the best option.


Internationale Hochschule Bad Honnef-Bonn GmbH


Manufacturing, Lean Manufacturing, ChatGPT