Hierarchical Criteria, Heuristic, Multiple criteria decision-making, FITradeoff, simulation.
This work proposes an approach to deal with multicriteria decision making problems in which the criteria are structured hierarchically, in Flexible and Interactive Tradeoff (FITradeoff). This approach uses partial preference information provided by the decision maker, based on a structured process within the scope of the Multiple Attribute Value Theory, to elicit the values of the scale constants. The approach is presented for the choice and ranking problematic and will be incorporated into the FITradeoff method, which is based on the traditional tradeoff procedure, axiomatically robust. The definition of this approach is made through the development of two mathematical models that will be compared to verify the similarity between them and the effectiveness in solving problems with hierarchical criteria. In addition, new concepts of hierarchy are incorporated into the models and two questioning heuristics are compared to verify which of them find the solution to the problem after fewer questions answered, aiming to reduce the cognitive effort of the decision maker. The models are distinguished because one of them uses ROC (Rank-Order Centroid) weights to determine the weights of the higher-level criteria and in the other the value of these weights is not known. Finally, the comparison between the two models is performed by running a simulation. The results obtained showed that the models are capable of solving problems with hierarchically structured criteria and that a new heuristic was needed to better adapt the approach. Furthermore, it was highlighted that the number of questions made to the decision maker is influenced by the number of families, distribution of weights of the lower-level criteria, number of lower-level criteria and number of alternatives.