Donghun Lee
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Donghun Lee has a total of 27 co-authors. Here are the top co-authors by number of shared publications:
- Kwanho Kim
Department of Industrial and Systems Engineering, Dongguk University, Seoul 04620, Republic of Korea
6 shared publications Ri Choe
2 shared publicationsSeokwoo Song
2 shared publications- Byungjun Bae
Electronics and Telecommunications Research Institute (ETRI), Daejeon 34129, Republic of Korea
1 shared publication Dongjin Lee
1 shared publication- Gyeong Rae Nam1 shared publication
Latest Publications
- Open Access
- 4 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 121630-121640
An Integrated Neural Network-Based Traffic Congestion Prediction for Material Handling Systems of Semiconductor Manufacturing
Rapid automated logistics within a factory are essential to maximize productivity. In semiconductor manufacturing, the most important logistics management is the efficient operation of overhead hoist transports (OHTs). To transfer wafers via OHTs without delays, it is necessary to predict short-term traffic congestion in the OHT railway accurately. However, the congestion prediction is a significant challenge due to the complexity of investigating all traffic conditions and dynamic traffic changes. Several studies have utilized machine learning approaches to address these concerns, but limitations arise in predicting the short-term congestion due to the performance bias stemming from large input features. Recurrent neural networks are effective in predicting traffic flow in transportation. However, they may not be suitable for OHT railway congestion prediction due to unpredictable loading/unloading events and varying traffic volumes. Therefore, this study proposes an integrated neural network-based method where multiple neural networks are trained considering current conditions of the railway network and expected changes in traffic conditions. To verify the effectiveness of the proposed method, a simulated dataset was used to reflecting real-world semiconductor fabrication. The experiment results demonstrate that the proposed method outperforms existing methods, including machine learning- and deep learning-based methods.
- 4 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 121630-121640
Open AccessAn Integrated Neural Network-Based Traffic Congestion Prediction for Material Handling Systems of Semiconductor Manufacturing
Rapid automated logistics within a factory are essential to maximize productivity. In semiconductor manufacturing, the most important logistics management is the efficient operation of overhead hoist transports (OHTs). To transfer wafers via OHTs without delays, it is necessary to predict short-term traffic congestion in the OHT railway accurately. However, the congestion prediction is a significant challenge due to the complexity of investigating all traffic conditions and dynamic traffic changes. Several studies have utilized machine learning approaches to address these concerns, but limitations arise in predicting the short-term congestion due to the performance bias stemming from large input features. Recurrent neural networks are effective in predicting traffic flow in transportation. However, they may not be suitable for OHT railway congestion prediction due to unpredictable loading/unloading events and varying traffic volumes. Therefore, this study proposes an integrated neural network-based method where multiple neural networks are trained considering current conditions of the railway network and expected changes in traffic conditions. To verify the effectiveness of the proposed method, a simulated dataset was used to reflecting real-world semiconductor fabrication. The experiment results demonstrate that the proposed method outperforms existing methods, including machine learning- and deep learning-based methods.
- Open Access
- 1 January 2025
- Published by IGI Global Scientific Publishing in Journal of Global Information Management
- Vol. 33 (1) , 1-25
Impact of Network Structures and Deep Learning on Financial Performance in Buyer-Supplier Networks
This study investigates the impact of network capabilities and deep learning techniques on companies' financial performance within buyer-supplier networks. It broadens the scope by incorporating network measures such as closeness and network constraint, whereas previous studies have primarily focused on suitable buyer-supplier relationships. The analysis evaluates the effects of these network measures on companies' financial performance metrics, including asset growth rate, return on assets, and more. In addition, this study explores the impact of extended networks on company performance using deep learning techniques. The results show that companies' network capabilities are positively associated with their financial performance, highlighting the critical role of network positions in driving success. Furthermore, the findings suggest that extending the network through deep learning offers significant benefits for companies.
- 1 January 2025
- Published by IGI Global Scientific Publishing in Journal of Global Information Management
- Vol. 33 (1) , 1-25
Open AccessImpact of Network Structures and Deep Learning on Financial Performance in Buyer-Supplier Networks
This study investigates the impact of network capabilities and deep learning techniques on companies' financial performance within buyer-supplier networks. It broadens the scope by incorporating network measures such as closeness and network constraint, whereas previous studies have primarily focused on suitable buyer-supplier relationships. The analysis evaluates the effects of these network measures on companies' financial performance metrics, including asset growth rate, return on assets, and more. In addition, this study explores the impact of extended networks on company performance using deep learning techniques. The results show that companies' network capabilities are positively associated with their financial performance, highlighting the critical role of network positions in driving success. Furthermore, the findings suggest that extending the network through deep learning offers significant benefits for companies.
- 31 August 2023
- Published by Korea Information and Communications Society in The Journal of Korean Institute of Communications and Information Sciences
- Vol. 48 (8) , 950-954
Image Captioning Network for Disaster Dataset Based on Transfer Learning and Disaster Convolution Block
Image captioning has been widely studied using various deep learning models on extensive and well-prepared datasets. Accurately captioning images of serious and sudden disasters is important; however, the study of disaster image captioning is yet to be thoroughly investigated compared to natural image captioning. Furthermore, existing image captioning models may need to perform better in generating captions for disaster images because there are fewer disaster images in popular datasets than non-disaster images. To address these problems, we refine and propose a cleaned disaster dataset and an image captioning model optimized for the dataset. Experimental results showed that our proposed model outperformed the existing model in terms of generating accurate captions for disaster images.
- 31 August 2023
- Published by Korea Information and Communications Society in The Journal of Korean Institute of Communications and Information Sciences
- Vol. 48 (8) , 950-954
Image Captioning Network for Disaster Dataset Based on Transfer Learning and Disaster Convolution Block
Image captioning has been widely studied using various deep learning models on extensive and well-prepared datasets. Accurately captioning images of serious and sudden disasters is important; however, the study of disaster image captioning is yet to be thoroughly investigated compared to natural image captioning. Furthermore, existing image captioning models may need to perform better in generating captions for disaster images because there are fewer disaster images in popular datasets than non-disaster images. To address these problems, we refine and propose a cleaned disaster dataset and an image captioning model optimized for the dataset. Experimental results showed that our proposed model outperformed the existing model in terms of generating accurate captions for disaster images.
- Open Access
- 9 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 58705-58714
Machine Learning-Based Day-Ahead Prediction of Price-Setting Scheduled Energy in the Korean Electricity Trading Mechanism
Power generation companies, which participate in the electricity trading mechanism, need to determine optimal capacity bidding strategy to maximize their operational profit in Korea. Since the Price-setting Power Generation Schedule determines the profit of power generators, it is important to predict Price-setting Scheduled Energy during the trading day before the bidding phase. We propose a methodology for predicting the Pricing-setting Scheduled Energy from the power exchange without optimizing it. Instead of simulating the planning process, machine learning algorithms are applied and compared in the process of predicting the Pricing-setting Scheduled Energy. The input variables consist of seasonal and price information including calendar, fuel cost, and system marginal price. Three categories of machine learning (ML) algorithms including single, bagging and boosting approaches are implemented and tested to compare their performances. The computational experiments show that ML algorithms with price variable are shown to be better in terms of the considered measures. In addition, boosting approach is more effective than single and bagging approaches.
- 9 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 58705-58714
Open AccessMachine Learning-Based Day-Ahead Prediction of Price-Setting Scheduled Energy in the Korean Electricity Trading Mechanism
Power generation companies, which participate in the electricity trading mechanism, need to determine optimal capacity bidding strategy to maximize their operational profit in Korea. Since the Price-setting Power Generation Schedule determines the profit of power generators, it is important to predict Price-setting Scheduled Energy during the trading day before the bidding phase. We propose a methodology for predicting the Pricing-setting Scheduled Energy from the power exchange without optimizing it. Instead of simulating the planning process, machine learning algorithms are applied and compared in the process of predicting the Pricing-setting Scheduled Energy. The input variables consist of seasonal and price information including calendar, fuel cost, and system marginal price. Three categories of machine learning (ML) algorithms including single, bagging and boosting approaches are implemented and tested to compare their performances. The computational experiments show that ML algorithms with price variable are shown to be better in terms of the considered measures. In addition, boosting approach is more effective than single and bagging approaches.
- Open Access
- 1 January 2022
- Published by Elsevier in SSRN Electronic Journal
Multi Deep Learning Based Machine Scheduler to Minimize Total Job Lateness by Considering the Dependency on Consecutive Allocations
This paper is concerned with a parallel dedicated machines scheduling (PDMS) problem with sequence-independent setup times to minimize total job lateness. The considered PDMS problem is challenging due to the allocation decisions considering the trade-off between setup and job lateness. For instance, when jobs with the same job types are repeatedly processed on a machine for minimizing setup, the job lateness is likely to be increased since its processing is completed after its due date. On the other hand, if dedicated machines are fully utilized to process jobs with multiply job types, setups tend to occur frequently, and those jobs are not processed in time eventually. In this paper, to solve this problem, we propose novel three machine schedulers based on deep neural network (DNN). Specifically, the proposed machine schedulers include the deep machine scheduler ( [[EQUATION]] ) and two enhanced versions of deep machine schedulers such as [[EQUATION]] and [[EQUATION]] They are designed to capture the hidden allocation patterns between jobs and machines by considering the dependency on the consecutive allocations in a machine. In particular, the proposed [[EQUATION]] and [[EQUATION]] additionally learn using the decision history and the random job selector by successfully preserving the previously acquired hidden allocation patterns from the trained schedulers. Through the extensive experiments, the proposed schedulers are successful to yield satisfactory scheduling performances. Moreover, in much more difficult scheduling settings, the suggested [[EQUATION]] obviously exhibits superior performances over the others.
- 1 January 2022
- Published by Elsevier in SSRN Electronic Journal
Open AccessMulti Deep Learning Based Machine Scheduler to Minimize Total Job Lateness by Considering the Dependency on Consecutive Allocations
This paper is concerned with a parallel dedicated machines scheduling (PDMS) problem with sequence-independent setup times to minimize total job lateness. The considered PDMS problem is challenging due to the allocation decisions considering the trade-off between setup and job lateness. For instance, when jobs with the same job types are repeatedly processed on a machine for minimizing setup, the job lateness is likely to be increased since its processing is completed after its due date. On the other hand, if dedicated machines are fully utilized to process jobs with multiply job types, setups tend to occur frequently, and those jobs are not processed in time eventually. In this paper, to solve this problem, we propose novel three machine schedulers based on deep neural network (DNN). Specifically, the proposed machine schedulers include the deep machine scheduler ( [[EQUATION]] ) and two enhanced versions of deep machine schedulers such as [[EQUATION]] and [[EQUATION]] They are designed to capture the hidden allocation patterns between jobs and machines by considering the dependency on the consecutive allocations in a machine. In particular, the proposed [[EQUATION]] and [[EQUATION]] additionally learn using the decision history and the random job selector by successfully preserving the previously acquired hidden allocation patterns from the trained schedulers. Through the extensive experiments, the proposed schedulers are successful to yield satisfactory scheduling performances. Moreover, in much more difficult scheduling settings, the suggested [[EQUATION]] obviously exhibits superior performances over the others.