A Comparative Study on Interval Estimation Methods of Process Capability Index Cpmk

• Main Authors: Hau-Wei Chen, 陳皓偉
• Other Authors: Chien-Wei Wu
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/74008620833695869148

碩士 === 逢甲大學 === 工業工程與系統管理學研究所 === 97 === Several process capability indices, including Cp, Cpk, Cpm, and Cpmk, have been proposed to provide numerical measures on manufacturing potential and actual performance. In particular, the index Cpmk is constructed by accounting for the process yield as well as the process loss. Unfortunately, the sampling distribution of the estimated Cpmk being complicated, this makes the exact confidence interval very difficult to establish. In this study, several existing approaches based on statistical theory and simulation techniques for constructing confidence intervals of the index Cpmk have been presented. The investigated methods include asymptotic distribution approach (AD), frequency distribution approach (FD), generalized confidence intervals approach (GCI), bootstrap approach and Bayesian approach (BA). A series of simulations is conducted to examine and compare the performance of these approaches in terms of coverage rate (CR) and the mean of lower confidence bound (MLCB). According to the simulation results, several conclusions and recommendations are made to provide simple guidelines for assessing process capability.