Papers by Author: James C. Chen

Paper TitlePage

Authors: Po Tsang B. Huang, James C. Chen, Yuan Tsan Jou
Abstract: The key element of the in-process surface roughness monitoring system is the decision-making model, which is utilized to analyze the input factors and then to generate a proper output. The success of the in-process monitoring system depends on the accuracy of the decision-making model. To increase the accuracy and reliability of model, it is important to reduce the variation of the inputs. To achieve this objective, an integration of regression and neural network was developed as a decision-making model in this research. In this integrated model, the regression model was applied as a filter to sort the input variables into groups. Furthermore, the grouped data was implemented to train and to generate different neural networks models to reduce the affection of input variation and increase the accuracy of the monitoring system. The input variables was first filtered by the threshold of regression model, and then analyzed by different neural networks model based on the filtered result. Finally, to evaluate the performance of the integrated model, the regression neural network and traditional neural networks were both developed for surface roughness monitoring system in an end milling operation to compare the accuracy of systems.
Authors: James C. Chen, Wun Hao Jaong, Cheng Ju Sun, Hung Yu Lee, Jenn Sheng Wu, Chung Chao Ku
Abstract: Resource-constrained multi-project scheduling problems (RCMPSP) consider precedence relationship among activities and the capacity constraints of multiple resources for multiple projects. RCMPSP are NP-hard due to these practical constraints indicating an exponential calculation time to reach optimal solution. In order to improve the speed and the performance of problem solving, heuristic approaches are widely applied to solve RCMPSP. This research proposes Hybrid Genetic Algorithm (HGA) and heuristic approach to solve RCMPSP with an objective to minimize the total tardiness. HGA is compared with three typical heuristics for RCMPSP: Maximum Total Work Content, Earliest Due Date, and Minimum Slack. Two typical RCMPSP from literature are used as a test bed for performance evaluation. The results demonstrate that HGA outperforms the three heuristic methods in term of the total tardiness.
Authors: James C. Chen, Shu Jen Hu, Yu Hsin Chen, C.L. Yang, Cheng Ju Sun, C.W. Chen
Abstract: Facility planning is crucial to the performance of array fabs in Thin Film Transistor - Liquid Crystal Display (TFT-LCD) industry. In order to avoid costly changes and modifications of fab layouts at the installation or production stages, the designers should carefully evaluate design alternatives and then select the best one during the design stage. A TFT-LCD manufacturing process consists of four basic processes: array, color filter, cell, and module. This research proposes an Array Fab Design Procedure (AFDP) to conduct quick calculations to develop and evaluate initial design alternatives for TFT-LCD array fabs. A series of practical formulae are presented to sequentially determine the related design parameters. The proposed AFDP provides a basis for rapid modeling and evaluation of array fab designs. AFDP is developed in Microsoft Visual Basic and data from real array fabs are used to demonstrate its effectiveness and efficiency. Results indicate that AFDP can quickly calculate the related fab design parameters. “Job shop with small bays” leads to the least average sheet moving distance. With AFDP as a tool, fab designers can evaluate design alternatives and conduct what-if analysis in the initial phase of fab design.
Authors: James C. Chen, Shu Jen Hu, Po Tsang B. Huang, Chiuhsiang Joe Lin, Kuo Jung Chao, Chih Cheng Chen
Abstract: A Finite Capacity Planning Policy (FCPP) is developed for multiple color filter (CF) fabs where each fab has several identical production lines. FCPP assigns customer orders to multiple fabs and multiple lines by taking into account each fab’s available capacity and Work-In-Process (WIP) level, as well as each order’s batch size, product type, process routing, and processing time. After all orders are assigned to fabs and lines, order release time and finish time are determined by the implementation of several modules. FCPP is developed in Microsoft Visual Basic for Application (VBA), and an AutoMod simulation model is also developed. A company with two CF fabs and two lines at each fab is treated as a case study, and industrial data from these fabs are collected and used to evaluate the performance of FCPP based on the design of experiments. Preliminary simulation results show that FCPP can effectively and efficiently balance the loading between fabs and also balance the loading between lines in each fab.
Showing 1 to 4 of 4 Paper Titles