Multivariate Nonparametric Control Chart With Storage Space

Currently, it is important to monitor multivariate data in a timely manner. The spatial-rank-based multivariate exponentially weighted moving average control chart (SREWMA) has better relative performance for multivariate data monitoring when the underlying distribution is non-normal. However, with the increasing amount of data, the computation requires increasing system memory. Thus, the running speed of the computer and monitoring efficiency of the SREWMA control chart are reduced. Regarding this issue, this paper provides the concept of storage space and adding storage space to the proposed SREWMA control chart. When the observations in storage space reach a fixed number of spatial-rank calculations, the computer begins to control the amount of calculations and memory. The SREWMA with storage space control chart greatly improves the running speed of the SREWMA control chart and is robust for various distributions. Finally, a real-data example from banknote authentication is provided to illustrate the proposed method.