Combining Numerical and Linguistic Information in Group Decision Making

People give information about their personal preferences in many different ways, depending on their background. This paper deals with group decision making problems in which the solution depends on information of a different nature, i.e., assuming that the experts express their preferences with numerical or linguistic values. The aim of this paper is to present a proposal for this problem. We introduce a fusion operator for numerical and linguistic information. This operator combines linguistic values (assessed in the same label set) with numerical ones (assessed in the interval [0,1]). It is based on two transformation methods between numerical and linguistic values, which are defined using the concept of the characteristic values proposed in this paper. Its application to group decision making problems is illustrated by means of a particular fusion operator guided by fuzzy majority. Considering that the experts express their opinions by means of fuzzy or linguistic preference relations, this operator is used to develop a choice process for the alternatives, allowing solutions to be obtained in line with the majority of the experts' opinions. © 1998 Elsevier Science Inc. All rights reserved.