Decision support system with Genetic algorithm for economic statistical design of Nonparametric control charts

A nonparametric control chart is a valuable alternative to a parametric control
chart when there is no parametric distribution assumption. Besides,
economic statistical design (ESD) is more practical than previous ones since
it considers not only cost optimization but also statistical constraints on
the ability of the chart to detect shifts in parameters. Motivated by these
advantages, we propose in this chapter an economic statistical design (ESD)
for two nonparametric control charts based on the sign and the Wilcoxon
signed-rank tests. The optimal parameters of the designed charts are found
through Decision Support System (DSS) approach with the genetic algorithm
(GA) method. The main advantages of the developed procedure are
that (1) it investigates constraints on both Type I and Type II errors; (2) it
reports a good behavior for detecting a shift with any out-of-control (OC)
distribution. The numerical experiments exhibit that the proposed charts
have very good performance over time in comparison with the classical ones
in the literature.