ANALYSIS OF COEFFICIENT VARIATION CONTROL CHART’S SIMULATION APPROACHED BY NORMAL AND CHI-SQUARE DISTRIBUTION AND T-STUDENT DISTRIBUTION

• Main Author: Safira, Jennifer
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/42475
• Institution: Institut Teknologi Bandung

The purpose of statistical process control (SPC) is to reduce variability. Usually SPC connected with constant mean and standard variation. But, there’re also processes that the mean is not constant. In these cases we need coefficent of variation. There are two methods to approach coefficient of variation’s statistic, using normal and chi-square distribution and using t-student distribution. This research use numeric simulation method to process different condition of X ?, S, and CV. The result show that coefficient of variation’s control chart at in control condition can be done by X ? and S’ control chart at out of control condition, for both of them. From the simulation of control limit value, for CV=0,2 the value of control limits are almost the same for all significance level. Furthermore there are some number of subgroup that have the same effective level for the same significance level. When the significance level getting smaller, the number of subgroup that have the same effective will get bigger. UCL and LCL for coefficient of variation using normal and chi-square distribution compare to coefficient of variation using t-student distribution the difference will get bigger when the significance level and number of subgroup getting bigger.