Abstract

TITLE :  PROPOSAL FOR THE IMPROVEMENT OF ATTRIBUTE CONTROL CHARTS USING DAUDIN'S METHODOLOGY

PRESENTED BY : Elena Pérez Bernabeu

DIRECTED BY : J. Manuel Jabaloyes Vivas and Andrés Carrión García

Control charts can be defined as a statistical tool used for controlling processes. It allows to the process 
responsible to distinguish between the reasons of variation that can appear: The variability owed to common 
reasons that allow us to predict the state of the process or the variability owed to special reasons, which 
do not allow us to predict the situation of the process in a certain moment.

We can classify Control charts, according to the characteristic to be controlled, in two big groups: 
Variable Control charts and Attribute Control charts.

It is stated that in the existing bibliography there are several studies about variable control charts, 
concretely, average control charts. On the other hand there are fewer contributions for the attribute 
control chart, which are beginning to be developed by some researchers in the last years. 

This Doctoral Thesis is centered in the proposal of the u control chart improvement. It is applied to this chart 
the Double Sampling designed by Daudin for the average chart in 1992. It is a particular case of the multiple sampling.
  
The main aim of this Doctoral Thesis is to know how Daudin's methodology is performed on the u control chart, and 
if in this cas, the power obtained is better than the one in the classic u chart.

This Doctoral Thesis structure consist of eight chapters and a part of annexes:

Chapter 1 : The precedents are introduced, a brief introduction to the attribute charts, as well as the aims 
            and the structure of this Doctoral Thesis.

Chapter 2 : A proposal of methodology is given for this Doctoral Thesis, and it is also explained the technologies 
            and the specific software.

Chapter 3 : It is defined the attribute charts, explaining with detail each of them, especially the u graph, on 
            which modifications have been applied in this Doctoral Thesis. Also there is done a revision of the 
            existing bibliography in this topic.

Chapter 4 : It is detailed an explanation of the methodology described in Daudin, 1992 and it is shown the 
            formulation of the problem for the u graph.

Chapter 5 : It is defined and explained the performance of a Genetic Algorithm. It has been used for calculating 
            the DS-U chart parameters that improve its power comparing to the u chart power.

Chapter 6 : The DS-U chart power in a value u0 is compared with the power of the classic u chart in the same value.

Chapter 7 : The behavior of the power is analyzed when it is tried to maximize the power of the DS-U chart 
            in an interval of values [u_{Inferior}, u_{Superior}] at the same time, in this case, those values 
            which remain next to the value of design u0, opposite to the power of the u chart in the same interval.

Chapter 8 : The conclusions obtained are summarized in this Doctoral Thesis, as well as the future research lines.

Annexes : It is shown all the information and charts referred to in chapters 6 and 7. It is also shown partially 
          the software source code used.