LAMP: LLM Assistant for Matrix Production

In the context of manufacturing, one of the key challenges related to the circular economy is the adaptation of production processes to enable highly customized products, to allow for small batches, and to increase sustainability. The matrix production paradigm can support this goal, as it sets the shop floor up as a series of independent modules. These can be reconfigured as necessary, adapting the use of resources as production requirements change. However, planning adaptations to processes in a matrix production system is currently done manually, which can be very time consuming and expensive.

To address this limitation, this project investigates the creation of an LLM-based assistant that can leverage existing knowledge to guide the adaptation of production processes. This includes shop floor information (e.g., specifications, process steps) and documentation on best practices for reconfiguration (e.g., guidelines, reports). Workers interact with the assistant to collaboratively plan and execute adaptations to the production process, which simplifies the task due to the assistant's intuitive natural language interface. At the same time, the assistant provides a direct and simple approach to knowledge transfer, as previous documents can be revisited when working on the process adaptation.