Montoring Schemes based on Cox-Snell Residuals

Multistage process monitoring with the aim of improving the product reliability is of main concern in today's competitive market. Diverse monitoring procedures have beendevised and implemented to monitor the reliability-related quality variables. The proposed schemes have addressed a singlestageprocess or the quality characteristics which are not dependent on the state of preceding quality variables. However,taking into account the relationship among quality measurementsis the major feature while monitoring multistage processes. In this paper, we propose two regression-adjusted control schemesbased on Cox-Snell residuals to effectively monitor a quality characteristic which is intended to evaluate the productreliability. Subsequently, the performance study has been conducted to calculate and compare the average run length (ARL) using Markov chain approach. The results reveal that theprocedure which uses cumulative sum (CUSUM) control chart outweighs its competing scheme. It is noteworthy that the proposed control procedures are applicable to service operations such as healthcare systems as well as manufacturing processes

Keywords:

Accelerated failure time (AFT) model , Multistage processes , Cox-Snell residuals , Cumulative sum (CUSUM) control chart , Cascade property , Markov chain