Robust regression metamodel for a maintenance float policy

Develops a robust maintenance float policy, that considers system design parameter settings that not only satisfy the system performance criteria but are also insensitive to various noise conditions. The experimental design strategy employed in the study involves the use of discrete event simulation. In the study, the strategy proposed involves solving a maintenance float policy using both the inner and outer arrays as advocated by Genichi Taguchi. Initial system variables and their parameter settings were chosen based on a prior study. These system variables were then classified into design factors and noise factors. An experimental design was developed using Taguchi’s orthogonal array, after which a simulation experiment was performed and additional data collected. Based on the results, regression was performed with the significant factors and interactions. From the regression analysis, a robust metamodel was developed. A cost model was also proposed.