Fuzzy logic and interval arithmetic based TOPSIS method for multi-attribute reverse auctions
Résumé
Most of the existing reverse auction methods are restricted to the price dimension and assume that the qualitative criteria are fixed prior to competitive source selection. Multicriteria reverse auctions (MCRAs) facilitate the negotiation of price and other qualitative and quantitative criteria. In this paper, the winner determination (WD) problem of MCRAs is formulated as a multicriteria decision-making problem. An extension of the TOPSIS (technique for order preference by similarity to ideal solution) method based on fuzzy logic and interval arithmetic is proposed to solve the WD problem where some attributes are qualitative and imprecise in nature and others are quantitative but difficult to represent in precise numerical values. In some cases, precise values are inadequate to model the criteria in real life. Moreover, because of the personalities of human beings, the decision is based less on information and more on personal judgments, resulting in a biased decision. Therefore, fuzzy linguistic variables are used here to map qualitative criteria, and at the same time intervals, data are used for quantitative preferences that are difficult to represent in exact numerical values. The correlation coefficient and standard deviation integrated method is used for automatic enumeration of the criteria weights, and a mechanism is also developed to determine the preferences of qualitative criteria using fuzzy linguistic variables. We present illustrative example to demonstrate the applicability of the proposed method.