A model for the decision-maker preferences in a polymer extrusion process

Abstract

The NN-DM is a method developed to find a mathematical model that represents the Decision-Maker (DM) by employing an artificial neural network (NN) in situations in which the preferences can be represented by a utility function. This paper presents further developments to the NN-DM method to find a model in a polymer extrusion process. The form of the DM's interaction, the domain assignment, the ranking process, and the performance assessment are adapted to a real context of a multi-objective optimization problem followed by a design decision. The DM is then requested to fill a matrix expressing his preferences considering pairwise comparisons expressing ordinal relations only. Two multi-objective optimization problems are tested, each one with three estimates of different Pareto-optimal fronts. The adapted NN-DM method is able to provide a model which sorts the available solutions from the best to the worst according to the DM's preferences.