B1608
Title: A new robust quality control chart
Authors: Didem Egemen - George Washington University (United States)
Baris Surucu - Orta Dogu Teknik Universitesi (Turkey) [presenting]
Abstract: Quality control charts are widely used in manufacturing industry to monitor characteristics of a process. The most popular of these control charts is due to Shewhart. Its derivatives and other control charts proposed in the literature are mainly based on simple random sampling and the assumption of normality. However, the normality assumption fails very much in practice for many data sets. A recent sampling method named as ranked set sampling is also known to be superior over simple random sampling in terms of its efficiency. We propose control charts for nonnormal symmetric distributions based on ranked set sampling. We construct quality control charts for short and long-tailed symmetric distributions under ranked set sampling scheme especially when parameters are unknown. We firstly make use of a robust estimation technique to estimate unknown parameters of the distributions. Secondly, we propose a new ranked set sampling scheme for the estimation of scale parameters. Then, we construct the control charts accordingly. Finally, we conduct a simulation study to show the performances of the newly proposed charts and also give real life examples to explain the subject.
Journal