Chen-Fu Chien

Data mining to improve personnel selection and enhance human capital: A case study in high-technology industry

The quality of human capital is crucial for high-tech companies to maintain competitive advantages in knowledge economy era. However, high-technology companies suffering from high turnover rates often find it hard to recruit the right talents. In addition to conventional human resource management approaches, there is an urgent need to develop effective personnel selection mechanism to find the talents who are the most suitable to their own organizations. This study aims to fill the gap by developing a data mining framework based on decision tree and association rules to generate useful rules for personnel selection. The results can provide decision rules relating personnel information with work performance and retention. An empirical study was conducted in a semiconductor company to support their hiring decision for indirect labors including engineers and managers with different job functions. The results demonstrated the practical viability of this approach. Moreover, based on discussions among domain experts and data miner, specific recruitment and human resource management strategies were created from the results.

Data mining for yield enhancement in semiconductor manufacturing and an empirical study

During wafer fabrication, process data, equipment data, and lot history will be automatically or semi-automatically recorded and accumulated in database for monitoring the process, diagnosing faults, and managing manufacturing. However, in high-tech industry such as semiconductor manufacturing, many factors that are interrelated affect the yield of fabricated wafers. Engineers who rely on personal domain knowledge cannot find possible root causes of defects rapidly and effectively. This study aims to develop a framework for data mining and knowledge discovery from database that consists of a Kruskal-Wallis test, K-means clustering, and the variance reduction splitting criterion to investigate the huge amount of semiconductor manufacturing data and infer possible causes of faults and manufacturing process variations. The extracted information and knowledge is helpful to engineers as a basis for trouble shooting and defect diagnosis. We validated this approach with an empirical study in a semiconductor foundry company in Taiwan and the results demonstrated the practical viability of this approach.

A Portfolio-Evaluation Framework for Selecting R&D Projects

This study aims to form the basis for constructing a framework for evaluating alternative portfolios of R&D projects. This study provides an extensive literature review on portfolio selection. Most of the existing studies deal with the portfolio selection problem by evaluating individual projects and then seeking ways to combine them for an R&D portfolio. However, the combination of individually good projects unnecessarily constitutes the optimal portfolio. In particular, this study discusses three portfolio effects: (1) the difference between the preference for the portfolio as a whole and the preference for the projects, (2) the interrelation among projects, (3) the size of portfolio selection problems. This study develops a three-phase framework for evaluating R&D portfolios and proposes a new taxonomy of the portfolio attributes (i.e. independent, interrelated, and synergistic). This study concludes with a discussion of future research, directed toward increasing the applicability of portfolio-selection approaches for managing R&D portfolios.

Using Bayesian network for fault location on distribution feeder

The Bayesian network is a probabilistic graphical model in which a problem is structured as a set of variables (parameters) and probabilistic relationships among them. The Bayesian network has been effectively used to incorporate expert knowledge and historical data for revising the prior belief in the light of new evidence in many fields. However, little research has been done to apply the Bayesian network for fault location in power delivery systems. We construct a Bayesian network on the basis of expert knowledge and historical data for fault diagnosis on a distribution feeder in Taiwan. The experimental results validate the practical viability of the proposed approach.

A DEA Study to Evaluate the Relative Efficiency and Investigate the District Reorganization of the Taiwan Power Company

In this study data envelopment analysis (DEA) models were applied to evaluate the relative efficiencies of twenty-two electricity distribution districts of the Taiwan Power Company (TPC) in Taiwan. The empirical study showed that the TPC districts have good overall efficiency. We found that eleven districts were inefficient. Most of the inefficient districts suffer from scale inefficiency to a greater degree than technical inefficiency. We suggested the specific improvement directions for the corresponding inefficient districts. This study also investigated district reorganization to increase the efficiency. The proposed district reorganization alternatives have higher efficiency scores than the current one.

An evolutionary approach to rehabilitation patient scheduling: A case study

Focusing on real settings, this study aimed to develop an evolutionary approach based on genetic algorithm for solving the problem of rehabilitation patient scheduling to increase service quality by reducing patient waiting time and improve operation efficiency by increasing the therapy equipment utilization. Indeed, due to partial precedence constraints of rehabilitation therapies, the problem can be structured as a hybrid shop scheduling problem that has received little attention to date. In addition, a mixed integer programming model was also constructed as a benchmark to validate the solution quality with small problems. Based on empirical data from a Medical Center in Taiwan, several experiments were conducted to estimate the validity of the proposed algorithm. The results showed that the proposed algorithm can reduce patient waiting time and enhance resource utilization and thus demonstrated the practicality of the proposed algorithm. Indeed, a decision support system embedded with the developed algorithm has been implemented in this medical center.

Manufacturing Intelligence to Exploit the Value of Production and Tool Data to Reduce Cycle Time

Cycle time reduction is crucial for semiconductor wafer fabrication companies to maintain competitive advantages as the semiconductor industry is becoming more dynamic and changing faster. According to Littles Law, while maintaining the same throughput level, the reduction in Work-in-Process (WIP) will result in cycle time reduction. On one hand, the existing queueing models for predicting the WIP of tool sets in wafer fabrication facilities (fab) have limitations in real settings. On the other hand, little research has been done to predict the WIP of tool sets with tool dedication and waiting time constraint so as to control the corresponding WIP levels of various tool sets to reduce cycle time without affecting throughput. This study aims to fill the gap by proposing a manufacturing intelligence (MI) approach based on neural networks (NNs) to exploit the value of the wealthy production data and tool data for predicting the WIP levels of the tool sets for cycle time reduction. To validate this approach, empirical data were collected and analyzed in a leading semiconductor company. The comparison results have shown practical viability of this approach. Furthermore, the proposed approach can identify and improve the critical input factors for reducing the WIP to reduce cycle time in a fab.

Using Rough Set Theory to Recruit and Retain High-Potential Talents for Semiconductor Manufacturing

To recruit and retain high-potential talent is critical for semiconductor companies to maintain competitive advantages in a modern knowledge-based economy. Conventional personnel selection methodologies focusing on static work and job analysis will no longer be appropriate for knowledge workers in high-tech industries. This paper aims to develop an effective data mining approach based on Rough Set Theory to explore and analyze human resource data for personnel selection and human capital enhancement. An empirical study was conducted in a leading semiconductor company in Taiwan to estimate the validity of the proposed approach for predicting work behaviors including performance and resignation. The results showed that latent knowledge can be discovered as a basis to derive specific recruitment and human resource management strategies. In particular, 29 rules have been adopted as references for recruiting the right talent. This paper concludes with discussions of empirical findings and future research directions.

Manufacturing intelligence to forecast and reduce semiconductor cycle time

Semiconductor manufacturing is one of the most complicated production processes with the challenges of dynamic job arrival, job re-circulation, shifting bottlenecks, and lengthy fabrication process. Owing to the lengthy wafer fabrication process, work in process (WIP) usually affects the cycle time and throughput in the semiconductor fabrication. As the applications of semiconductor have reached the era of consumer electronics, time to market has played an increasingly critical role in maintaining a competitive advantage for a semiconductor company. Many past studies have explored how to reduce the time of scheduling and dispatching in the production cycle. Focusing on real settings, this study aims to develop a manufacturing intelligence approach by integrating Gauss-Newton regression method and back-propagation neural network as basic model to forecast the cycle time of the production line, where WIP, capacity, utilization, average layers, and throughput are rendered as input factors for indentifying effective rules to control the levels of the corresponding factors as well as reduce the cycle time. Additionally, it develops an adaptive model for rapid response to change of production line status. To evaluate the validity of this approach, we conducted an empirical study on the demand change and production dynamics in a semiconductor foundry in Hsinchu Science Park. The approach proved to be successful in improving forecast accuracy and realigning the desired levels of throughput in production lines to reduce the cycle time.

A novel timetabling algorithm for a furnace process for semiconductor fabrication with constrained waiting and frequency-based setups

This study aims to solve the scheduling problem arising from oxide–nitride–oxide (ONO) stacked film fabrication in semiconductor manufacturing. This problem is characterized by waiting time constraints, frequency-based setups, and capacity preoccupation. To the best of our knowledge, none of the existing studies has addressed constrained waiting time and frequency-based setups at the same time. To fill this gap, this study develops a genetic algorithm for batch sequencing combined with a novel timetabling algorithm. For validation, we conducted several experiments based on empirical data. As a benchmark for small-sized problem instances, a mixed-integer linear programming model was used. The results show that the proposed algorithm optimally solves most cases of the ONO scheduling problem in real settings and significantly outperforms dispatching rule-based heuristics.