This paper purposes to modify multivariate control chart with multivariate spatial signs and ranks. The Monte Carlo simulation is used to compare the performance of the control chart based on the ARL. The modified control chart is sensitive to small shifts monitoring in the process mean vector and provide fast signals for detecting small shifts with more efficiency for symmetric t-distribution data. The results show that the SSRM outperforms MEWMA, dMEWMA, and SSRdM control charts in detecting small shifts of all smoothing parameter and in detecting moderate and large shifts outperform with the large smoothing parameter. The most industrial has current data and the industrial situations prefer large values of the smoothing parameter, thus the SSRM is more appropriate to kurtosis data in detecting small shifts for process mean monitoring and not interest in the resolution of data.
Modified Multivariate Control Chart Using Spatial Signs and Ranks for Monitoring Process Mean: A Case of t-Distribution
4 Views
7 Downloads