A dynamic plate design problem via reinforcement learning.
Saved in:
| Title: | A dynamic plate design problem via reinforcement learning. |
|---|---|
| Authors: | Shi, Fengyuan1,2 (AUTHOR), Meng, Ying1,2 (AUTHOR) mengying@ise.neu.edu.cn, Tang, Lixin1 (AUTHOR), Zhao, Shengnan1,2 (AUTHOR) |
| Source: | International Journal of Production Research. Jun2026, Vol. 64 Issue 11, p4606-4634. 29p. |
| Subjects: | Reinforcement learning, Cutting stock problem, Steel industry, Markov processes, Resource management, Mathematical optimization |
| Abstract: | In the steel industry, plate design plays a crucial role in improving resource utilisation. The plate design problem can be viewed as a variant of the two-dimensional cutting stock problem, which is known to be NP-hard. This paper investigates a dynamic plate design problem arising in steel production, incorporating several practical features, including rolling deformation, maximum and minimum length restrictions, and dynamically arriving orders. These features significantly increase the complexity of the problem. In this paper, we formulate the dynamic plate design problem as a Markov decision process with supermodularity. Then, the monotonic relationship between the optimal action set and the system state is analysed. Accordingly, we propose a reinforcement learning approach with the lightweight network, where an action space compression strategy is introduced to reduce the size of the decision space without sacrificing solution quality. Computational experiments show that the proposed approach can solve problem instances within 35 seconds, satisfying the efficiency requirements of real-world production environments. Compared with other existing methods, the proposed approach achieves reductions of up to 16.543% in the number of plates used and 231.16% in trim loss. [ABSTRACT FROM AUTHOR] |
| Copyright of International Journal of Production Research is the property of Taylor & Francis Ltd 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: | In the steel industry, plate design plays a crucial role in improving resource utilisation. The plate design problem can be viewed as a variant of the two-dimensional cutting stock problem, which is known to be NP-hard. This paper investigates a dynamic plate design problem arising in steel production, incorporating several practical features, including rolling deformation, maximum and minimum length restrictions, and dynamically arriving orders. These features significantly increase the complexity of the problem. In this paper, we formulate the dynamic plate design problem as a Markov decision process with supermodularity. Then, the monotonic relationship between the optimal action set and the system state is analysed. Accordingly, we propose a reinforcement learning approach with the lightweight network, where an action space compression strategy is introduced to reduce the size of the decision space without sacrificing solution quality. Computational experiments show that the proposed approach can solve problem instances within 35 seconds, satisfying the efficiency requirements of real-world production environments. Compared with other existing methods, the proposed approach achieves reductions of up to 16.543% in the number of plates used and 231.16% in trim loss. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 00207543 |
| DOI: | 10.1080/00207543.2026.2626535 |
