A Bayesian analysis of reliability improvement experiments. (6th September 2017)

• Record Type: Journal Article
• Authors: Yu, I‐Tang
• Is Part Of: Quality and reliability engineering international. Volume 33:Number 8(2017:Dec.)
• Page Start: 2531
• Page End: 2538
• Publication Date: 2017-09-06
• Subjects: Bayesian estimation -- designed experiment -- log‐location‐scale model -- Markov chain Monte Carlo; Reliability (Engineering) -- Periodicals; Quality control -- Periodicals; High technology -- Periodicals; 620.00452
• DOI: 10.1002/qre.2215
• Languages: English
• ISSNs: 0748-8017

Abstract

Abstract: The number of effects can be studied in a log‐location‐scale regression model used in analyzing a reliability improvement experiment is restricted to the number of runs, which is usually small. In many real examples, only the main effects and a few 2‐factor interactions are considered. In this work, we propose using a Bayesian approach to analyze reliability improvement experiments. By specifying a prior on the effects, the number of effects can be studied is no longer restricted to the number of runs, and aliased effects can all be identified and estimated simultaneously. We analyze 2 real data sets to demonstrate the proposed approach. The results show that when complex interactions are present, the proposed approach can provide a more reliable result.