COMPSTAT 2016: Start Registration
View Submission - COMPSTAT
A0584
Title: Web computing of robust methods in Acceptance Sampling for Weibull variables Authors:  Miguel Casquilho - IST, University of Lisbon (Portugal) [presenting]
Elisabete Carolino - ESTeSL-IPL (Portugal)
M Rosario Ramos - FCiencias.ID (Portugal)
Isabel Barao - Faculty of Sciences - University of Lisbon (Portugal)
Abstract: Acceptance sampling (AS) is used to inspect the input or output of a process, mainly in manufacturing. A sampling plan is designed to determine a procedure that, if applied to a series of lots of a given quality and based on sampling information, leads to a specified risk of accepting or rejecting the lots. Classic AS by variables assumes Gaussian distribution, as treated in industry standards, which is sometimes an abusive assumption leading to wrong decisions. AS for variables with asymmetric and/or heavy tailed distributions is then a relevant topic. Specific AS plans are derived for the case of the Weibull distribution. As an alternative to these AS plans for the Weibull, traditional plans are used with robust estimators. The estimators are the sample median for location and a modified version of the sample standard deviation and the Total Range for scale estimates. The problem of determining AS plans by variables is addressed for the Weibull distribution with unknown parameters. The aim is to apply computing over the Web, through an application made available to any user, needing no software installation, in order to conduct a simulation study on these methods, about classical plans, specific plans and plans using the robust estimates for location and scale