Improve Nonparametric EWMA Sign Charts Based on EPQ Model.

Traditional control chart applications to monitor process shifts are not suitable for the situations where observations come from a non-normal or unknown distribution. In order to trace a small-shift quality deviation of a process a special quality control chart is required. We have developed an economic-statistical design of a new nonparametric exponentially weighted moving average (EWMA) sign chart for improving the detection capabilities of process stability and tracing small-shift quality deviation. The contribution of our research is twofold: (1) an economic-statistical model of the nonparametric EWMA sign chart with a new economic production quantity (EPQ) design that enables to minimize the total cost of a small-shift quality deviations; (2) model of a dynamic adjusted selector of the crossover and mutation rate in GA's evolutions for the parameter optimization of EWMA sign chart design. [ABSTRACT FROM AUTHOR]