O método de Equação Integral com Quadratura Gaussiana para otimizar os parâmetros do gráfico de controle multivariado de Somas Acumuladas
The fast growth of data technologies acquisition and the use of computers have increased the interest in techniques of statistical emergent control, such as applied multivariate control charts for monitoring processes that involve the simultaneous control of the quality characteristics or process va...
Guardado en:
| Autor principal: | |
|---|---|
| Otros Autores: | |
| Formato: | text (thesis) |
| Lenguaje: | por |
| Publicado: |
Universidade Federal de Santa Catarina (Brasil)
2009
|
| Materias: | |
| Acceso en línea: | https://dialnet.unirioja.es/servlet/oaites?codigo=20854 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | The fast growth of data technologies acquisition and the use of computers have increased the interest in techniques of statistical emergent control, such as applied multivariate control charts for monitoring processes that involve the simultaneous control of the quality characteristics or process variables correlated. In this work the method of Integral Equation with Gaussian Quadrature to optimize the main parameters ARL, k and h of the multivariate cumulative sum control chart (MCUSUM) is proposed. This method involves the derivation of an analytical integral equation, whose numerical resolution way Gaussian Quadrature allows the user to obtain the approximate solution of these parameters using a computer program developed from this method. Furthermore, from this methodology, a systematic approach to design a statistical chart of MCUSUM optimal control based on these parameters, which minimizes the average cost of operation, is also proposed. This, guides the user to decide what size to change the vector of means is important to detect, and taking this as a base, selecting those parameters MCUSUM great chart that address statistical issues such as minimizing the number of false alarms and maximizing the ability of detection of real changes. The results of this study reveal the method of Integral Equation proposed as an excellent alternative to optimize the parameters of the chart MCUSUM to be a more versatile procedure that provides better results for the value of ARL with higher speed of calculation compared with the method of simulations and the relative simplicity of implementation in a computer program |
|---|