Premium
A consensus reaching process for large‐scale group decision making with heterogeneous preference information
Author(s) -
Wu Zheng,
Liao Huchang
Publication year - 2021
Publication title -
international journal of intelligent systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.291
H-Index - 87
eISSN - 1098-111X
pISSN - 0884-8173
DOI - 10.1002/int.22469
Subject(s) - preference , preference elicitation , group decision making , aggregate (composite) , ordinal scale , computer science , aggregation problem , process (computing) , ordinal regression , representation (politics) , selection (genetic algorithm) , scale (ratio) , preference learning , homogeneous , information retrieval , artificial intelligence , machine learning , mathematics , statistics , mathematical economics , psychology , geography , social psychology , materials science , cartography , combinatorics , politics , political science , law , composite material , operating system
Many group decision making (GDM) models enable experts to use only one preference information representation form. It is natural to allow experts to express preferences in various formats considering the heterogeneity of experts. In this case, how to reach the consensus of a group from heterogeneous preference information is an attractive research issue. This study proposes a consensus reaching process for large‐scale GDM with heterogeneous preference information. First, we review various preference formats including preference orderings, numerical assessments, interval‐valued assessments, and linguistic assessments. To facilitate the heterogeneous information aggregation, we classify experts into subgroups according to their preference types rather than the similarities of preference values, and then aggregate the homogeneous preference values in each subgroup. The subgroup priorities derived by homogeneous methods are then aggregated into global priorities. An ordinal consensus measuring process based on individual orderings is introduced. To reach the ordinal consensus, optimization models are constructed to ensure each subgroup's preferences equivalent to the global preferences, and the recommended ranges and strength of preference modification are given to experts. Finally, the proposed method is validated by an illustrative example about blockchain platform selection.