Fuzzy weights estimation method based on the linguistic expression of criteria relevance

Abstract

In this article, the quantified expression of criteria weights in decision-making processes is considered. A decision maker is often able to express his or her preferences only linguistically. The concept of fuzzy numbers, or linguistic variables, is used in order to state these vague preferences. A fuzzy number is considered as one with a triangular membership function, which renders an approximate value of the weight of a particular criterion according to linguistic terms about its importance. To set the strict (crisp) value of weights, the optimization model based on the maximin principle is proposed in two modifications. The proposed approaches are compared with other methods. Finally, the introduced methods are applied to a rather illustrative but real case of investment decision making in the capital market. The main results are discussed.