A hesitant fuzzy perceptual-based approach to model linguistic assessments

• Autores: Olga Porro Martorell
• Directores de la Tesis: Mónica Sánchez Soler (dir. tes.) , Núria Agell Jané (codir. tes.)
• Lectura: En la Universitat Politècnica de Catalunya (UPC) ( España ) en 2021
• Idioma: español
• Texto completo no disponible (Saber más ...)
• Resumen

  • Multiple-criteria or multiple-attribute group decision-making is a sub-field of operations research that seek to find a common and representative solution given the preferences elicited by a pre-defined group, over a set of alternatives and with respect to a set of coherent criteria (or attributes). Recently, the modelling of natural language in these processes has captured the attention of many researchers. Most of the evaluations in a group-decision making context are inherently imprecise, incomplete or vague, and therefore, experts feel more comfortable using their language rather than numerical values.

    The use of hesitant fuzzy linguistic term sets is one of the recent tools that enables the modelling of linguistic assessments in multiple-criteria decision-making. Nonetheless, advances in hesitant linguistic multi-attribute group decision making require the development of structures flexible enough to deal with unbalanced and multi-granular linguistic information. More tools are needed in order to really grasp the differences in the qualitative reasoning processes of each individual.

    This thesis, firstly, introduces a perceptual-based distance able to capture differences between unbalanced linguistic assessments, which is based on a lattice structure of hesitant fuzzy linguistic terms. Secondly, this distance is used to define a perceptual-based centroid or central opinion which, in turn, is used to define a consensus measure or degree of agreement within the group. Thirdly, with the aim to deal with multi-perceptual group decision-making contexts, where each decision maker has its own qualitative reasoning approach, a perceptual-based transformation function and a projected algebraic structure are defined. The developed tools can deal with different multi-granularity linguistic environments.

    Two applications are presented to demonstrate the utility, relevancy and feasibility of the methods. On the one hand, a specific perceptual-based classification and ranking method is introduced and applied to a real group decision making problem in an educational setting. This framework is used to classify and rank a set of secondary students according to their degree of entrepreneurial competency, which is based on real data provided by the Andorra Government. On the other hand, an extended fuzzy multi-perceptual linguistic TOPSIS is designed and applied to a real group decision making problem in the context of smart city governance. This perceptual extension is used to assess the criteria governing the strategic decision making process of energy multinational companies when deciding where to expand its sustainable services and products.