Essays on imprecise representation of uncertainty considering evidence theory and FUZZY

Abstract

This paper contains the compendium of three articles on the representation of uncertainty. The two first ones deal with the classical theory of evidence or Dempster - Shafer theory. In these articles, the classification of conflict in the theory of evidence is approached, since bodies of evidence with high level of conflict lead to counterintuitive results when the Dempster Rule of Combination is used. In the first article, the conflict classification is made from class profiles by using the ELECTRE TRI method. In the second article, the classification is made from reference alternatives of each conflict class, the classification of conflict having been obtained through the outranking flow of the PROMETHEE method among the reference alternatives and the alternatives to be classified. In both articles, the parameters of the two methods are obtained by means of disaggregation approaches where the parameters are generated from alternatives pre-classified in conflict in the first article and reference alternatives in the second. Finally, the third article deals with the financial analysis in the development of a new product where a real options model is added by using dynamic programming, the modeling of uncertainty through FUZZY triangular numbers. Thus, in this model it is possible to consider the different types of uncertainty contained in the development of new products.