Alpha Error of Taguchi Method with Different OAs for NTB Type QCH by Simulation

Abstract

Taguchi method has been widely used for parameter design in many industrial applications. Nevertheless, it has been the subject of discussion and much debate in different platforms. This research proposes an extension to ongoing research by investigating the alpha error of Taguchi method with two-, three-, and four-level orthogonal arrays (OAs) for the nominal-the-best (NTB) type quality characteristic (QCH) type via simulation. With each array, it is assumed that QCH values are normally distributed with the same mean and standard deviation. Consequently, the null hypothesis that all factors are insignificant is true. The alternative hypothesis is that at least one factor is identified as significant. Simulation is conducted for 10 cycles each of 10,000 runs. The results showed that the alpha error is very high, which indicates that insignificant factors are misidentified as significant with high probability. In practice, this may provide misleading conclusions about parameter design. In conclusion, Taguchi’s quality engineering concepts are of great importance. However, his method is found inefficient for parameter design.

Key Words:

• Alpha error
• orthogonal arrays
• simulation
• Taguchi method

Additional information

Notes on contributors

Abbas Al-Refaie

Ming-Hsien Caleb Li is a Professor in the Department of Industrial Engineering and Systems Management at Feng Chia University, Taiwan. His interests are Six Sigma Management, Quality Engineering, Taguchi Method, Design of Experiments, and Statistical Quality Control. He is a member of Chinese IIE and Chinese Society for Quality.

Ming-Hsien Li

Abbas Al-Refaie is an Assistant Professor in the Department of Industrial Engineering at University of Jordan, Amman. His research interests include Data Envelopment Analysis, Robust Design, Statistical Quality Control, Design of Experiments, Taguchi Methods, Operation Research and Optimization, and Quality Management.