Function Based Adaptive Ewma Control Chart to Monitor Multivariate Process Dispersion

The adaptive EWMA (AEWMA) charts have been broadly perceived as an effective process monitoring instrument due to their capability to rapidly detect the parameter(s) shifts of diverse sizes. In this paper, a function based AEWMA control chart is designed to efficiently observe the stability of the variance-covariance matrix of a multivariate normally distributed process when individual observations are composed. The idea is to assess the shift using a biased free estimator founded on the EWMA statistic and after that adapt the smoothing constant of plotting EWMA statistic through the proposed continuous function. It is shown in the light of the Monte Carlo simulation studies that the proposed AEWMA chart ended up being proficient in identifying variation in the covariance matrix of the process as matched to the existing EWMA and AEWMA charts. As an example, data from the bimetal thermostat industry is used to demonstrate the applicability of the proposed AEWMA chart