Μέθοδοι ερμηνείας του εκτός ελέγχου σήματος πολυμεταβλητών διαγραμμάτων για τη διασπορά της διεργασίας
Interpereting the out-of control signal of a multivariate dispersion control chart
View/ Open
Subject
Έλεγχος διεργασιών -- Στατιστικές μέθοδοι ; Ποιοτικός έλεγχος -- Στατιστικές μέθοδοι ; Παραγωγή -- Διοίκηση και οργάνωση -- Ποιοτικός έλεγχος ; Πολυμεταβλητή ανάλυση ; Process control -- Statistical methods ; Quality control -- Statistical methods ; Production management -- Quality control -- Statistical methods ; Multivariate analysisAbstract
Statistical process control is a widely used tool which aims in controlling and improving an industrial or manufacturing process. Univariate techniques can be used for monitoring one characteristic in a process but in many situations more characteristics must be monitored simultaneously. By monitoring the characteristics independently, will lead on false conclusions because the relation between the variables is not taken into account. So, the usage of multivariate statistical process control is a necessity for avoiding such situations. While most techniques developed in the multivariate case deal with monitoring the mean vector of the process, monitoring the dispersion is equally important due to the fact that by controlling the dispersion, the mean will fluctuate less leading in a more stable process. The present dissertation initially provides an introduction in statistical process control. Next, the simple case of univariately monitoring the dispersion is presented. In addition an extensive literature review on multivariate control charts for the dispersion has been made and finally, we present a comparison of several multivariate control charts that appear in literature.