Information granulation of linguistic information as a basis for improving consensus in group decision making

Abstract

Group decision making involves a number of agents verbalizing their testimonies about a collection of options in order to rank them from best to worst. To provide their opinions, the agents feel more comfortable using linguistic values (or words) instead of numbers because words are usually used by humans to interact with others. However, when words are used, they have to be made operational to be fully utilized. To do so, we propose in this study an information granulation of the linguistic information expressed as an optimization task in which the consensus level achieved among the assessments expressed by the group of agents is maximized by an appropriate association of the words on information granules that are formed as intervals.