Abstract:
Simple Random Sampling (SRS) is the usual data collection method used for constructing control charts. In this paper, a method based on Ranked-Set Sampling (RSS) is developed to construct a control chart for the mean ( -RSS chart) and for the median (MRSS chart). The method uses RSS to construct a 100(1-α)% confidence interval for the population mean and then this confidence interval is used to construct the control charts. Furthermore, a new method of data collection is proposed that we call Median-Set Sampling (MSS). The data from MMS can be used to construct a control chart for mean ( -MSS chart) and a control chart for median (MMSS chart). The control charts constructed using SRS, MSS and MMSS are compared. We have found that if the process mean does not shift then the average run length (ARL) of the MRSS control chart was better than the ARL of the other charts. If there is a small shift in process mean then the MMSS chart was found to be the best method for detecting the shift. If there is a large shift of process mean, then the -RSS chart, MRSS chart, -MSS chart and MMSS chart were found to show similar performance for detecting the shift.