ADJUSTMENT OF FORECASTS IN MULTICRITERIA ANALYSIS BASED ON PARAMETERIZABLE FUNCTIONS

This paper proposes improvements in the multicriteria method of decision support TODIM, based on the prospect theory, by means of a parameterizable function. This approach aims to make the method more flexible and more adherent to the actual behaviour of decision-makers. Empirical validation was carried out with data from 20 higher education students, analyzing their preferences and weights assigned to criteria and alternatives. The results obtained with the original phi function are compared to those of the new parameterizable function, evaluating its impact on the prediction of choices. The proposal aims to bring greater robustness to the descriptive and predictive capacity of TODIM in contexts of multicriteria decision at risk.