Large-area OLED substrate printing path planning method based on multi-head GAT imitation learning to solve partitioned integer programming

Abstract Inkjet printing is considered a very promising technology in the field of organic light-emitting diode (OLED) substrate manufacturing. Compared with the vapor deposition process, inkjet printing has the advantages of a simple process, high material utilization and applicability to a wide ra...

Full description

Saved in:
Bibliographic Details
Main Authors: Jiacong Xiong, Jiankui Chen, Yiqun Li, Xiao Yue, Yu Fu, Zhouping Yin
Format: Article
Language:English
Published: Nature Portfolio 2025-07-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-08355-x
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849335289703563264
author Jiacong Xiong
Jiankui Chen
Yiqun Li
Xiao Yue
Yu Fu
Zhouping Yin
author_facet Jiacong Xiong
Jiankui Chen
Yiqun Li
Xiao Yue
Yu Fu
Zhouping Yin
author_sort Jiacong Xiong
collection DOAJ
description Abstract Inkjet printing is considered a very promising technology in the field of organic light-emitting diode (OLED) substrate manufacturing. Compared with the vapor deposition process, inkjet printing has the advantages of a simple process, high material utilization and applicability to a wide range of display manufacturing processes. However, during the inkjet manufacturing process, the resolution of the printhead (nozzle per inch, NPI) usually does not match the pixel resolution of the substrate (pixel per inch, PPI). Therefore, the travel path of the printhead module must be planned to minimize the number of print cycles required to complete the pattern. This involves a challenging multi-objective optimization process. The difficulty intensifies in large-area OLED production, where angular alignment errors in the substrate are magnified. This results in an exponential increase in pixels requiring planning, with the total pixel pit count reaching hundreds of millions. In addition, the time complexity of the planning problem grows exponentially, denoted as O(mn), and the space complexity grows rapidly with the matrix dimension. This problem is NP-hard. This problem has a significant impact on the productivity of the manufacturing process. In this paper, a large-area substrate printing planning algorithm based on graph attention networks and integer programming (GIP-LASP) is established. GIP-LASP provides partitioning rules and parallel modeling methods specifically for substrate misalignment angles, and proposes imitation learning based on a multi-head graph attention network on a SCIP solver, which is applied to the solution of the printing planning problem. The planning method was implemented on a G4.5 half-size substrate with a resolution of 394 PPI, and the color filter (CF) layer was successfully printed.
format Article
id doaj-art-9bf04db76d9a4ebfaf6bd955606257ea
institution Kabale University
issn 2045-2322
language English
publishDate 2025-07-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-9bf04db76d9a4ebfaf6bd955606257ea2025-08-20T03:45:19ZengNature PortfolioScientific Reports2045-23222025-07-0115112210.1038/s41598-025-08355-xLarge-area OLED substrate printing path planning method based on multi-head GAT imitation learning to solve partitioned integer programmingJiacong Xiong0Jiankui Chen1Yiqun Li2Xiao Yue3Yu Fu4Zhouping Yin5The State Key Laboratory of Intelligent Manufacturing Equipment and Technology, School of Mechanical Science and Engineering, Huazhong University of Science and TechnologyThe State Key Laboratory of Intelligent Manufacturing Equipment and Technology, School of Mechanical Science and Engineering, Huazhong University of Science and TechnologyThe State Key Laboratory of Intelligent Manufacturing Equipment and Technology, School of Mechanical Science and Engineering, Huazhong University of Science and TechnologyThe State Key Laboratory of Intelligent Manufacturing Equipment and Technology, School of Mechanical Science and Engineering, Huazhong University of Science and TechnologyThe State Key Laboratory of Intelligent Manufacturing Equipment and Technology, School of Mechanical Science and Engineering, Huazhong University of Science and TechnologyThe State Key Laboratory of Intelligent Manufacturing Equipment and Technology, School of Mechanical Science and Engineering, Huazhong University of Science and TechnologyAbstract Inkjet printing is considered a very promising technology in the field of organic light-emitting diode (OLED) substrate manufacturing. Compared with the vapor deposition process, inkjet printing has the advantages of a simple process, high material utilization and applicability to a wide range of display manufacturing processes. However, during the inkjet manufacturing process, the resolution of the printhead (nozzle per inch, NPI) usually does not match the pixel resolution of the substrate (pixel per inch, PPI). Therefore, the travel path of the printhead module must be planned to minimize the number of print cycles required to complete the pattern. This involves a challenging multi-objective optimization process. The difficulty intensifies in large-area OLED production, where angular alignment errors in the substrate are magnified. This results in an exponential increase in pixels requiring planning, with the total pixel pit count reaching hundreds of millions. In addition, the time complexity of the planning problem grows exponentially, denoted as O(mn), and the space complexity grows rapidly with the matrix dimension. This problem is NP-hard. This problem has a significant impact on the productivity of the manufacturing process. In this paper, a large-area substrate printing planning algorithm based on graph attention networks and integer programming (GIP-LASP) is established. GIP-LASP provides partitioning rules and parallel modeling methods specifically for substrate misalignment angles, and proposes imitation learning based on a multi-head graph attention network on a SCIP solver, which is applied to the solution of the printing planning problem. The planning method was implemented on a G4.5 half-size substrate with a resolution of 394 PPI, and the color filter (CF) layer was successfully printed.https://doi.org/10.1038/s41598-025-08355-xPrinting displaySubstrate angular deflectionInteger programmingParallel modellingMulti-head graph attention network
spellingShingle Jiacong Xiong
Jiankui Chen
Yiqun Li
Xiao Yue
Yu Fu
Zhouping Yin
Large-area OLED substrate printing path planning method based on multi-head GAT imitation learning to solve partitioned integer programming
Scientific Reports
Printing display
Substrate angular deflection
Integer programming
Parallel modelling
Multi-head graph attention network
title Large-area OLED substrate printing path planning method based on multi-head GAT imitation learning to solve partitioned integer programming
title_full Large-area OLED substrate printing path planning method based on multi-head GAT imitation learning to solve partitioned integer programming
title_fullStr Large-area OLED substrate printing path planning method based on multi-head GAT imitation learning to solve partitioned integer programming
title_full_unstemmed Large-area OLED substrate printing path planning method based on multi-head GAT imitation learning to solve partitioned integer programming
title_short Large-area OLED substrate printing path planning method based on multi-head GAT imitation learning to solve partitioned integer programming
title_sort large area oled substrate printing path planning method based on multi head gat imitation learning to solve partitioned integer programming
topic Printing display
Substrate angular deflection
Integer programming
Parallel modelling
Multi-head graph attention network
url https://doi.org/10.1038/s41598-025-08355-x
work_keys_str_mv AT jiacongxiong largeareaoledsubstrateprintingpathplanningmethodbasedonmultiheadgatimitationlearningtosolvepartitionedintegerprogramming
AT jiankuichen largeareaoledsubstrateprintingpathplanningmethodbasedonmultiheadgatimitationlearningtosolvepartitionedintegerprogramming
AT yiqunli largeareaoledsubstrateprintingpathplanningmethodbasedonmultiheadgatimitationlearningtosolvepartitionedintegerprogramming
AT xiaoyue largeareaoledsubstrateprintingpathplanningmethodbasedonmultiheadgatimitationlearningtosolvepartitionedintegerprogramming
AT yufu largeareaoledsubstrateprintingpathplanningmethodbasedonmultiheadgatimitationlearningtosolvepartitionedintegerprogramming
AT zhoupingyin largeareaoledsubstrateprintingpathplanningmethodbasedonmultiheadgatimitationlearningtosolvepartitionedintegerprogramming