ON BAYESIAN ESTIMATIONS AND MUTI-CHARACTERISTICS FOR PROCESS CAPABILITY INDICES

• Main Authors: Hsin-Lin Kuo, 郭信霖
• Other Authors: LIANG-YUH OUYANG
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/67075429663540750593

博士 === 淡江大學 === 管理科學學系 === 91 === The study of the process capability indices correctly measures the inherent variability of the process, and is devoid of extraneous disturbing influences when the process is in statistical control. It is a vital part of an overall quality improvement project. In this dissertation, first we consider the problem of estimations of indices with μ≠T, with μ = m, and Cpp with μ≠T under Bayesian framework. Several priors such as Non-informative prior, Gamma prior and Weibull hazard prior are considered for deriving the Bayesian estimators and lower bound of the credible intervals under some conditions. It provides a more general situations and suits for engineers and practical applications. Bothe( 1992 ) pointed out that a customer purchases an entire product, consisting of many characteristics. Currently, most of the methods exist for measuring the process capability of a single characteristic of a product. It’s extremely difficult to measure the process ability to just via those measures, which are only for individual characteristics. Hence, we develop a check-list and an evaluated procedure of integrated analysis for the process capability of an entire product, and provide the continual improvement of the process environment. Six Sigma approach concentrates on clarifying business strategy and making sure that everything relates to company objectives. It is vital to investigate each part of the business in such a way everyone can understand the causes of variation that can be led to improvement in both processes and performance. In this study, we demonstrate the match index ( Cp, Cpk ) in the 6σ procedure to decide whether a given process meets the capability requirement under quality control. Furthermore, we also provide the other indices matches, ( Cp, Cpm ), ( Cp, Cpmk ) and ( Cp, Cpp ), which are in more general situations suitable for engineers to apply.