Control Chart for Heavy Tailed Distributions

ABSTRACT Standard control charts with control limits determined by the mean and standard error of the mean are constructed based on the assumption that the distribution of the quality characteristic being monitored follows a normal distribution. However, this assumption is not always valid. It is proposed to use a chart based on computing the control limits using the process mean and the standard error of the least absolute deviation for the case where the process quality characteristics follow a heavy tailed t distribution. Such a control chart is more effective than the normal distribution based chart since it has a low out-of-control average run length for both small and large values of process shift. OPSOMMING 'n Kontrolekaart wat gebruik maak van kontrolelimiete gebaseer op die standaardafwyking van die geringste absolute limiet word ontwerp vir 'n t-verdeling met 'n betekenisvolle stert. Simulasietoetse vir vergelyking van die voorgestelde kontrolekaart met normaalverdeelde kontrolekaarte toon dat korter gemiddelde looplengtes voor diagnose van beheerverlies uitgewys word, bereik word. 10