ARL computation and parameters optimization for MEWMA control chart ased on the Markov chain

Average run length (ARL) is one of the important indices in performance evaluation on quality control charts, the research on which will set up the foundation of quality control chart parameters optimization, performance evaluation and application. An ARL computation model based on the Markov chain was studied for the multivariate exponentially weighted moving average (MEWMA) control chart, the Matlab program was designed and implemented. The convergence of the algorithm and the performance of the MEWMA control chart in different conditions through the analysis of the computation results were analyzed. A procedure for parameters optimization of MEWMA control chart was proposed. An computation example and some optimization results was presented.