Attribute weighting model based on a new inconsistency index

Abstract

This work was proposed as the first objective to demonstrate that a consistency index proposed in 2015 as part of an attribute weighting model is not objective enough. The second objective is the proposal of a new optimization model to weight attributes that is based on an idea similar to that of the analyzed model where the limitations detected in it are not evident. To achieve the above, the significance and scope of the consistency and inconsistency indices proposed in the previous work from where the limitation was induced in the first one is corroborated with the analysis of a case study. The new model is based on extended fuzzy logic techniques and on new concepts introduced by the authors such as the weighted generalized inconsistency index of order q and the internal and external confidence coefficients.