Group screening and Taguchi design in the optimization of multi-echelon maintenance float simulation metamodels

This paper presents a regression model for a multi-echelon maintenance float policy. The model is based on the use of group screening and the Taguchi approach in determining the experimental design plans. Subsequently, a simulation model is built and the resulting outputs analyzed. Regression metamodels are developed for the average equipment utilization. This model is operational within the bounds defined for the independent variables. The proposed approach enhances the use of simulation in solving complex production problems. Additionally, it reduces the need to study every level of the independent variable. Therefore, when the number of independent variables is large, it is better to identify the common attributes of these variables in relation to the dependent variable. They can then be grouped into classes and used as the new factors in the design.