Search Articles

Home / Articles

EWMA-TYPE CHART BASED ON REGRESSION ESTIMATOR USING AUXILIARY VARIABLE AND THE GENERALIZED LIKELIHOOD RATIO TEST STATISTIC

. Samson Offorma Ugwu, Everestus Okafor Ossai, Tobias Ejiofor Ugah, Emmanuel Ikechukwu Mba, Michael Chinonso Eze, Felix Obi Ohanuba, Precious Ndidiamaka Ezra and Nnamdi Michael Nwakobi


Abstract

Statistical process control (SPC) is an important integral aspect of every manufacturing process whose aim is to maintain product or service standard. Control charts are the most indispensable aspect of the SPC because of their statistical background. The cumulative sum (CUSUM) and the exponentially weighted moving average (EWMA) charts are better options to the Shewhart -charts owing to their abilities to timely detect small and moderate shifts in the process parameter. To increase the sensitivity of the EWMA chart, the use of auxiliary variables in the estimation of its charting statistic has been proposed. Also, to make the chart monitor for both the location and dispersion parameters of the process concurrently, its charting statistic has been modified by using the generalized likelihood ratio statistic. This work proposes a generalized likelihood ratio statistic which would use auxiliary variable that is based on regression estimation technique for joint monitoring of the process parameters. The average run length (ARL) and the median run length (MRL) of the chart were evaluated in simulation and the ARLs compared with those of the rival control charts in the literature. It was noticed that the proposed chart outperformed rival charts in detecting assignable causes of variations in the process.

 

Keywords: auxiliary information; average run length; control charts; generalized ratio statistic.

Download :