REMEDIAL APPROACHES TO DECREASE THE EFFECT OF MEASUREMENT ERRORS ON SIMPLE LINEAR PROFILE MONITORING

In most profile monitoring applications, the explanatory variables are not fixed from profile to profile. However, in most studies, they are assumed to be fixed values. Furthermore, the observed units usually contain some source of uncertainty referred as "measurement errors." In this paper, first the effect of neglecting the measurement errors effect on detection capability of two common control charts for Phase II monitoring of simple linear profiles is evaluated. Then, three remedial approaches including ranked set sampling, multiple measurement and increasing sample size are utilized to decrease the mentioned effect. Simulation studies in terms of average run length (ARL) metric show that neglecting the measurement errors adversely affects the capability of both charts. The results also confirm that the remedial approaches adequately compensate for the mentioned effect.