Title: Robust estimation of the process dispersion for standard deviation control charts
Authors: Shih-Chou Kao
Addresses: Department of Business Administration, Kao Yuan University, No. 1821, Jhongshan Rd., Lujhu District, Kaohsiung City 821, Taiwan
Abstract: The detection ability of control charts varies with the robustness of estimators against contamination. The aim of this current study is to develop six types of estimators based on the scale A method with distributed weight functions and compare the performance of various dispersion control charts based on these functions under normal and contaminated normal environments. The values of their mean squared error are compared to those of existing estimators in diffuse and localised disturbances. The process-monitoring abilities of phase II control charts using phase I contaminated estimators are assessed using disturbances and process shifts. The estimator with the logistic distributed weight function performs the best against disturbances, with its average run lengths being closer to those in uncontaminated cases compared to other estimators. [Submitted: 15 December 2019; Accepted: 21 February 2022]
Keywords: average run length; ARL; mean squared error; MSE; scale A estimator; trimming; weight function.
European Journal of Industrial Engineering, 2023 Vol.17 No.3, pp.408 - 430
Accepted: 21 Feb 2022
Published online: 01 May 2023 *