Group Decision Making Models Based On Multi-Granular Information
Loading...
Date
2015-01
Authors
FENG, ZHANG
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
A variety of information encapsulated in different representations is used by Decision Makers (DMs) to express their preferences during the decision making process. All such expressions (e.g., ordinal, linguistic, fuzzy, and interval) constitute information with different granularity which is useful to capture uncertainty, vagueness, imprecision in DMs’ opinions, preferences, or assessments in an effective manner. Based on widely use of multi-granular information in group decision making (GDM) problems, a large number of GDM models have been proposed in the literature. Although these GDM models provide a valuable means to handle multi-granularity information, there are still some drawbacks associated with different GDM models that need improvements. The purpose of this thesis is to provide some new GDM models and related methods for the improvement of existing GDM models from the viewpoint of multi-granular information. Specifically, GDM models based on four types of information, namely ordinal, linguistic, fuzzy, and interval, are investigated in this research. For the ordinal preferences, a two-stage dynamic GDM model based on a consensus reaching process is proposed, where a Power Average (PA) operator is employed to aggregate ordinal information with consideration of the relationship (e.g. agreement or disagreement) among the DMs. Additionally, a data cleaning process is integrated into the consensus reaching process to evade the bias caused by the conflicting opinions.
Owing to the internal linear ordinal property of linguistic term sets, a supplementary method of linguistic GDM models is obtained by extending the Dominant-based Rough Set Approach on support function to linguistic information. Since the group decisions generated
by fuzzy GDM models are typically represented by fuzzy numbers, and the collective group decisions generated by intervals GDM models are normally represented by interval matrices, two new methods capable of handling fuzzy and interval information are then proposed to derive the final ranking from the collective group decisions. The main contribution and innovation of this thesis lies in the improvements of various GDM models with multi-granular information and related methods. Mathematical proofs are presented, and efficiency of the resulting GDM models is demonstrated by using a number of examples and case studies.
Description
Keywords
Group Decision Making Models Based On , On Multi-Granular Information