FlexSim-Simulated PCB Assembly Line Optimization Using Deep Q-Network

The balance scheduling of Printed Circuit Board (PCB) assembly lines plays a crucial role in enhancing production efficiency. Traditional scheduling methods rely on fixed heuristic rules, which lack flexibility and adaptability to changing production demands. To address this issue, this paper propos...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Engineering proceedings Jg. 75; H. 1; S. 34
Hauptverfasser: Jinhao Du, Jabir Mumtaz, Wenxi Zhao, Jian Huang
Format: Journal Article
Sprache:Englisch
Veröffentlicht: MDPI AG 01.10.2024
Schlagworte:
ISSN:2673-4591
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The balance scheduling of Printed Circuit Board (PCB) assembly lines plays a crucial role in enhancing production efficiency. Traditional scheduling methods rely on fixed heuristic rules, which lack flexibility and adaptability to changing production demands. To address this issue, this paper proposes a PCB assembly line scheduling method based on Deep Q-Network (DQN). The PCB assembly line model is constructed using the FlexSim simulation tool, and the optimal scheduling strategy is learned through the DQN algorithm. Comparative analysis is conducted against traditional heuristic rules. Experimental results indicate that the DQN-based scheduling method achieves substantial improvements in balance and production efficiency. For instance 1, the DQN approach achieved a total completion time (S) of 2.521 × 105, compared to the best heuristic rule result of 2.541 × 105. Similarly, for instance 2 and instance 3, the DQN method achieved total completion times of 2.549 × 105 and 2.522 × 105, respectively, outperforming all heuristic rules evaluated. This study provides a novel approach and method for intelligent scheduling of PCB assembly lines.
ISSN:2673-4591
DOI:10.3390/engproc2024075034