NUMERICAL ANALYSIS OF CONTROL CHART FOR AUTOCORRELATED DATA BASED ON AVERAGE RUN LENGTH
A basic assumption of the control chart is that the data are independent and identically distributed (iid). Sampling techniques on the controlling process so quickly can lead to an autocorrelation. In some industrial processes autocorrelation is attached to the process itself, so it can not be avoid...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/29403 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | A basic assumption of the control chart is that the data are independent and identically distributed (iid). Sampling techniques on the controlling process so quickly can lead to an autocorrelation. In some industrial processes autocorrelation is attached to the process itself, so it can not be avoided. One of the parameters that determine the performance characteristics of the control chart is the Average Run Length (ARL). To see the effect of autocorrelation on the performance of the control chart based on ARL, a simulation was performed to calculate the ARL value in the X Shewhart, X Residue, EWMA, EWMA Residue, and ARMA control charts. The applied data is the ARMA process (1,1) with different values of autocorrelation. Based on the ARL values generated from the simulation it can be concluded that autocorrelation has an effect on the performance of the control chart. The ARL values on each control chart vary depending on the magnitude of the autocorrelation value. At the final conclusion of the analysis results on the performance of the control chart based on the ARL value, it is found that each control chart has its own advantages and disadvantages when applied to data that has autocorrelation. |
|---|