الفهرس Only 14 pages are availabe for public view
 |  | from | 134 |  |  |
| | | from | 134 | | |
Abstract
Control chart is still one of the most important tools of statistical process control. The emergent need to build computer-integrated manufacturing (CIM) systems has created the necessity to automate the operation of control charts, and to utilize the information of the chart more effectively. In this direction artificial neural network (NN) technology, which try to emulate the massively parallel and distributed processing of the brain, is suggested as an alternative to traditional quality control charts. Pervious researches focus on the development of univariate NN models. This research aims at developing NN approaches that can be used in statistical process control (SPC) instead of the traditional quality control charts. The developed approaches try to detect the shift in the process mean and the process variability (standard deviation). Two backpropagation NN models were designed. One of them was trained by the single variable quality control data and the other by the bivariate quality control data. Effectiveness of the NN models is evaluated through the number of wrong decisions (Type I and Type II errors) and the in control and out-of-control average run lengths. Results of comparing NN models to the control charts show that the proposed NN models work successfully in the univariate and bivariate cases.
KEYWORDS: Control charts, Neural Networks, Univariate NN model, Bivariate NN model..