Abstract
In the modern manufacturing environment, the ability to collect and refine data in real time to deliver high-quality data is increasingly important in maintaining a competitive advantage and operational efficiency. This paper proposes a conceptual architectural framework for the continuous knowledge updating of Retrieval-Augmented Generation-Large Language Model-based systems in a smart factory environment. The proposed framework provides theoretical models and validation methodologies, laying the groundwork for future practical implementations. Existing Retrieval-Augmented Generation-Large Language Model systems rely on static knowledge bases that are not able to effectively reflect new information in a real-time, changing manufacturing environment. The proposed framework design uses a data stream processing layer, a data integration layer, and a continuous learning layer as core components; in particular, the knowledge integration layer provides a mechanism for the efficient processing of real-time data and continuous learning. This study is significant in that it presents a mathematical model and a systematic verification methodology that can quantitatively predict the performance and scalability of the proposed architecture, thus providing practical design guidance for the implementation of Retrieval-Augmented Generation-Large Language Model systems in smart factory environments. This paper is organized as follows: Materials and Methods provides a detailed description of the architecture and methodology. Theoretical Analysis and Discussion covers the theoretical analysis and discussion, including performance prediction models, validation methodologies, and their practical implications. Finally, Conclusions summarizes the research findings and outlines directions for future work.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call