Comparison of Bivariate Copulas on The Hotelling’s T2 and Mewma Control Charts
The aim of this paper proposes three types of copulas on the Hotelling’s T2 and MEWMA control charts. Observations are from an exponential distribution which Monte Carlo simulation based on copula function. We compare the performance of control charts by the Average Run Length (ARL). Dependence between random variables are measured by Kendall’s tau. The results show that the Gumbel copula of MEWMA control chart can be used for almost all shifts.
Index Terms— copulas, Monte Carlo simulation, Hotelling’s T2, MEWMA.