Optimization of buffer design for mixed-model sequential production line based on simulation and reinforcement learning.
Saved in:
| Title: | Optimization of buffer design for mixed-model sequential production line based on simulation and reinforcement learning. |
|---|---|
| Authors: | Choi, Jonghwan1 (AUTHOR) lgm3@g.skku.edu, Park, Jisoo1 (AUTHOR) jisoo7589@g.skku.edu, Noh, Sang Do1 (AUTHOR) sdnoh@skku.edu, Lee, Ju Yeon2 (AUTHOR) jylee@seoultech.ac.kr |
| Source: | Journal of Intelligent Manufacturing. Dec2025, Vol. 36 Issue 8, p5695-5714. 20p. |
| Subjects: | Reinforcement learning, Buffer storage (Computer science), Mass customization, Manufacturing processes, Assembly line methods, Resource allocation, Industrial efficiency, Computer simulation |
| Abstract: | Recently, as the market environment changes rapidly and customer demands diversify, the manufacturing paradigm is shifting towards mass customization and personalization. Consequently, companies are striving to establish optimal production systems that emphasize flexibility and efficiency. In particular, sequential production lines utilizing several machines have recently transitioned to small-batch production, particularly for the manufacture of automobiles and printed circuit boards (PCBs). In the context of mixed-model sequential production lines, production processes become complicated and uncertain, resulting in various challenges, such as varying processing times for each machine based on the product and setup times for machines when products change. A production buffer between machines can serve as an effective solution to these challenges by enhancing efficiency and productivity through improved material flow between sequential production processes. However, production lines often face constraints in terms of available space for buffer allocation, and the associated costs must also be considered. Therefore, it is essential to adopt a Buffer Allocation Problem (BAP) method that accounts for these factors. This paper proposes a simulation and reinforcement learning-based buffer optimization method designed to derive the optimal number, size, and location of buffers for mixed-model sequential production lines while considering both spatial and cost constraints. The proposed method's system framework is presented, with defined components, including a reinforcement learning module for optimal buffer information and a discrete event simulation module to assess rewards in the learning process. The optimization method is validated through application in a real-world manufacturing site, presented as a case study. [ABSTRACT FROM AUTHOR] |
| Copyright of Journal of Intelligent Manufacturing is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Recently, as the market environment changes rapidly and customer demands diversify, the manufacturing paradigm is shifting towards mass customization and personalization. Consequently, companies are striving to establish optimal production systems that emphasize flexibility and efficiency. In particular, sequential production lines utilizing several machines have recently transitioned to small-batch production, particularly for the manufacture of automobiles and printed circuit boards (PCBs). In the context of mixed-model sequential production lines, production processes become complicated and uncertain, resulting in various challenges, such as varying processing times for each machine based on the product and setup times for machines when products change. A production buffer between machines can serve as an effective solution to these challenges by enhancing efficiency and productivity through improved material flow between sequential production processes. However, production lines often face constraints in terms of available space for buffer allocation, and the associated costs must also be considered. Therefore, it is essential to adopt a Buffer Allocation Problem (BAP) method that accounts for these factors. This paper proposes a simulation and reinforcement learning-based buffer optimization method designed to derive the optimal number, size, and location of buffers for mixed-model sequential production lines while considering both spatial and cost constraints. The proposed method's system framework is presented, with defined components, including a reinforcement learning module for optimal buffer information and a discrete event simulation module to assess rewards in the learning process. The optimization method is validated through application in a real-world manufacturing site, presented as a case study. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 09565515 |
| DOI: | 10.1007/s10845-024-02525-w |
