TY - GEN
T1 - A Study on the Performances of the SPRT Control Chart When Estimating Process Parameters
AU - Teoh, Jing Wei
AU - Teoh, Wei Lin
AU - El-Ghandour, Laila
AU - Chong, Zhi Lin
AU - Teh, Sin Yin
PY - 2023/4/27
Y1 - 2023/4/27
N2 - The sequential-probability-ratio-test (SPRT) control chart is well-known for its outstanding detection ability towards a wide range of process mean shifts. In the existing literature, the SPRT chart has been regularly designed from the standpoint that the process parameters are known. Nonetheless, industrial practitioners reveal that these process parameters are hardly known. They often estimate these process parameters using a modest number of Phase-I samples. This in turn leads to different outcomes obtained across practitioners. In this paper, we present detailed insights about the distribution of the conditional ATS when the process parameters are estimated. Also, some statistical properties of the ATS through Monte Carlo simulation are presented in this paper. Particularly, the standard deviation of the ATS is examined in this paper to evaluate the practitioner-to-practitioner variability. Results show that a large Phase-I sample, and a large desired maximum mean shift in optimization design, are required to reduce volatilities in the average and standard deviation of the ATS values. Finally, we provide recommendations for the Phase-I sample sizes tailored to specific objectives of the SPRT chart when the process parameters are estimated.
AB - The sequential-probability-ratio-test (SPRT) control chart is well-known for its outstanding detection ability towards a wide range of process mean shifts. In the existing literature, the SPRT chart has been regularly designed from the standpoint that the process parameters are known. Nonetheless, industrial practitioners reveal that these process parameters are hardly known. They often estimate these process parameters using a modest number of Phase-I samples. This in turn leads to different outcomes obtained across practitioners. In this paper, we present detailed insights about the distribution of the conditional ATS when the process parameters are estimated. Also, some statistical properties of the ATS through Monte Carlo simulation are presented in this paper. Particularly, the standard deviation of the ATS is examined in this paper to evaluate the practitioner-to-practitioner variability. Results show that a large Phase-I sample, and a large desired maximum mean shift in optimization design, are required to reduce volatilities in the average and standard deviation of the ATS values. Finally, we provide recommendations for the Phase-I sample sizes tailored to specific objectives of the SPRT chart when the process parameters are estimated.
KW - Average time to signal
KW - Estimated parameters
KW - Quality control
KW - SPRT control chart
KW - Standard deviation of the average time to signal
UR - http://www.scopus.com/inward/record.url?scp=85161380498&partnerID=8YFLogxK
U2 - 10.1007/978-981-19-8406-8_36
DO - 10.1007/978-981-19-8406-8_36
M3 - Conference contribution
AN - SCOPUS:85161380498
SN - 9789811984051
T3 - Lecture Notes in Electrical Engineering
SP - 449
EP - 463
BT - Proceedings of the 9th International Conference on Computational Science and Technology. ICCST 2022
A2 - Kang, Dae-Ki
A2 - Alfred, Rayner
A2 - Ismail, Zamhar Iswandono Bin Awang
A2 - Baharum, Aslina
A2 - Thiruchelvam, Vinesh
PB - Springer
T2 - 9th International Conference on Computational Science and Technology 2022
Y2 - 27 August 2022 through 28 August 2022
ER -