Studies on the Effects of Estimator Selection in Robust Parameter Design under Asymmetric Conditions

摘要: The primary goal of robust parameter design (RPD) is to determine the optimum operating conditions that achieve process performance targets while minimizing variability in results. To this goal, typical approaches RPD problems use ordinary least squares methods obtain response functions for mean and variance by assuming experimental data follow a normal distribution are relatively free contaminants or outliers. Consequently, most common estimators used initial tier estimation sample variance, as they very good when these assumptions hold. However, it often case such assumed do not exist practice; notably, inherent asymmetry pervades system outputs. If unaccounted for, can affect results tremendously causing quality estimates obtained using standard deviation deteriorate. Focusing on asymmetric conditions, paper examines several highly efficient alternatives deviation. We then incorporate into modeling optimization ascertain which tend yield better solutions skewness exists. Monte Carlo simulation numerical studies substantiate compare proposed with traditional approach. Copyright © 2012 John Wiley & Sons, Ltd.