Abstract
In inspection of manufactured parts inaccuracies of the measurement process compel to use test limits which are more strict than the given specification limits. To avoid violation of the prescribed bound γ on the consumer loss, in practice overly conservative informal approaches are used leading to a loss of yield. In this paper approximations are derived for which γ is realized with high precision. The expressions involved are sufficiently simple and explicit to allow extension to the more realistic case where the parameters of the underlying distributions need to be estimated. Various criteria are presented to measure the closeness of the resulting stochastic consumer losses to γ. Some attention is devoted to the sampling schemes involved.
Similar content being viewed by others
References
Albers W, Kallenberg WCM (1991) A simple approximation to the bivariate normal distribution with large correlation coefficient. Technical Report 1011, University of Twente. To appear in Journal of Multivariate Analysis
Albers W, Kallenberg WCM, Otten GD (1992) Accurate test limits with estimated parameters. Technical Report 1024, University of Twente. To appear in Technometrics
Cox DR, Wermuth N (1991) A simple approximation for bivariate and trivariate normal integrals. International Statistical Review 59:263–269
Easterling RC, Johnson ME, Bement TR, Nachtsheim CJ (1991) Statistical tolerancing based on consumer's risk considerations. Journal of Quality Technology 23:1–11
Kotz S, Johnson NL (1985) Encyclopedia of statistical sciences, Volume 5 Wiley New York
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Albers, W., Kallenberg, W.C.M. & Otten, G.D. Setting test limits under prescribed consumer loss. Metrika 41, 163–181 (1994). https://doi.org/10.1007/BF01895315
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF01895315