• distribution-free;
  • nonparametric procedure;
  • self-starting;
  • spatial rank;
  • multivariate EWMA;
  • robustness;
  • statistical process control


Nonparametric control charts are useful in statistical process control when there is a lack of or limited knowledge about the underlying process distribution, especially when the process measurement is multivariate. This article develops a new multivariate self-starting methodology for monitoring location parameters. It is based on adapting the multivariate spatial rank to on-line sequential monitoring. The weighted version of the rank-based test is used to formulate the charting statistic by incorporating the exponentially weighted moving average control scheme. It is robust to non-normally distributed data, easy to construct, fast to compute and also very efficient in detecting multivariate process shifts, especially small or moderate shifts which occur when the process distribution is heavy-tailed or skewed. As it avoids the need for a lengthy data-gathering step before charting and it does not require knowledge of the underlying distribution, the proposed control chart is particularly useful in start-up or short-run situations. A real-data example from white wine production processes shows that it performs quite well. © 2012 Wiley Periodicals, Inc. Naval Research Logistics 59: 91–110, 2012