Page 187 - Special Topic Session (STS) - Volume 3
P. 187

STS539 Muhammad Abid et al.
                          14
                Yang et al.  observed that the in-control ARL results of EWMA NS  chart is
            not consistent when  differs from 0.5. So, to overcome this deficiency, Yang
                 11
            et al.  also developed the non-parametric arcsine EWMA (EWMA     NAS ) chart
                                                               +
                                                              
            using the arcsine transformation i.e.,  =  −1 (√ ). The distribution of 
                                                              
                                                                                     1
            follows the normal distribution with mean =  −1 (√) and variance = ( )
                                                                                    4
            (cf. Yang et al. ).
                          11
            The plotting statistic of EWMA NAS  chart is written as:
                                         =  + (1 − )   −1               (3)
                                                  
                Firstly,    0  =  −1 (√) and  the ( ) =  −1 (√) and  the
                                                                
                                1                             11
            ( ) =  ( ), respectively (cf. Yang et al. ).
                           2− 4
            So, the control limits of the EWMA NAS  chart are:
                                                                 1
                                    =  −1 (√) + √    ( ),
                                                  2− 4
                                            =  −1 (√) ,
                                          
                                                                 1
                                                             
                                                               ( ).
                                    =  −1 (√) − √ 2− 4
            2.3 Non-parametric sign CUSUM control chart
                                                                 12
                Using  the  statistic  given  in  (1)  Yang  and  Cheng   developed  the  two
            plotting  statistic  i.e.,    +   and    −   of  the  non-parametric  CUSUM  sign
            (CUSUMNS) chart as follows:
                                  +
                                               +
                                                     +
                                 = (0,  −1  +  − ( + )) }                           (4)
                                  
                                                     
                                                            0
                                                                 +
                                              −
                                  −
                                 = (0,  −1  − ( − ) +  )
                                                       0
                                                                
                                  
            where  = 1, 2, . .. and initially,  = 0 and  = 0.  The statistics given in (6)
                                           +
                                                       −
                                           
                                                       
            are plotted against their control limits ℎ and −ℎ, respectively. The process is
            considered to be out-of-control if  ≥ ℎ or  ≤ −ℎ, else, it is in-control.
                                                         −
                                               +
                                                         
                                               
            2.4 Non-parametric mixed control chart
                             13
                Abbasi  et  al.   proposed  a  mixed  nonparametric  EWMA  and  CUSUM
            (MECNAS)  control  charts  using  the  arcsine  transformation.  The  statistic  of
            MECNAS control chart is given below:
                                     +
                                                                    +
                                 = max [0, ( −  ) −  +  −1  }      (9)
                                    
                                                 
                                                      0
                                                            
                                                                     −
                                   _
                                = max [0, −( −  ) −  + + −1
                                                 
                                                            
                                                      0
                                   
            2.5 Proposed arcsine HWMA control chart
                                                                                      2
                   14
            Hunter  observed that the plotting statistic of EWMA chart proposed by Page
            is given more weights to the current observations and fewer weights to the
                                                                         32
            previous  observations.  To  overcome  this  deficiency,  Abbas   proposed  a
            HWMA chart and its statistic is defined as:
                                             =  ̅ + (1 − ) ̿ −1                                       (5)
                                                   
                                             
            where  ̅  represent the sample mean for ℎ sample,  ̿ −1  is the mean of the means of
                    
            previous  − 1 samples  and  is the sensitivity parameter lies between zero and one
            i.e., 0 <  ≤ 1. The  ̿ −1  is also defined as:
                                                               176 |I S I   W S C   2 0 1 9
   182   183   184   185   186   187   188   189   190   191   192