文摘
This research develops two alternatives to the classical statistical control charts, a robust alternative that improves the behaviour of these control charts in the phase I and one new proposal for the autocorrelated processes following a multivariate optic by means of the state-space model. Firstly we analyse the main classical control charts (from an univariate and a multivariate optic). Then, we study the problems related to the necessity of iterative calculations in the adjustment of a control chart during phase I. As well as the problems related to the breach of the independence supposition (in phase II). Like contribution, we think about the use of robust estimators called "Trimmed Means" to make agile the adjustment and estimation of the parameters of the control chart in phase I. In addition, we propose the application of the state space model. Later of modelization we set out the use of two control charts; A control chart for the innovations that depicts variations of great magnitude and cuts temporary duration and another for the states that depicts variations in the process general behaviour, variations of smaller intensity and greater temporary duration. Finally, we analyse the behaviour of these methods in a practical application. We use the developed method to their application on the cutler sector in Albacete, a sector of great importance for the economy in Albacete.;Keywords. Statistical Quality Control, Robust estimators, Multivariate Time Series, State- Space Model.