Comparative analysis of Phase II autocorrelated simple linear profile monitoring methods with estimated parameters.

AbstractThere are Phase II monitoring schemes for linear profiles with between-profile autocorrelation that assume known parameters. However, their performance under parameter estimation has not been evaluated. This study fills this gap by assessing Phase II monitoring schemes for autocorrelated simple linear profiles with estimated parameters and recommending Phase I sample sizes. We evaluate four schemes: T2, EWMA/R, and EWMA3 proposed by Noorossana, Amiri, and Soleimani and VEWMA3 proposed by Wang and Huwang. Simulations show that as the number of in-control Phase I profiles increases, the in-control ARL approaches the target, and the Standard Error of ARL decreases for all schemes. VEWMA3 is the most effective in detecting parameter shifts but requires a larger dataset of in-control profiles to stabilize performance, especially at higher autocorrelation. This study provides practical Phase I sample size guidelines to enhance statistical process control reliability. [ABSTRACT FROM AUTHOR]