Determination of the fuzzy weights of criteria in MADM

Abstract

One of the most important difficulties of researchers in multi attribute decision making problems is the determination of weights of criteria. Different methods are introduced for this purpose in MADM literature. Some methods are based on individual ideas of experts or decision makers and some others are based on the structure of decision matrix. The existing methods that are introduced for weighting of different criteria suffer from the lack of precisions of weightings given by individual and incapability to use the concept of pairwise comparison of criteria. In this paper a new procedure is introduced to determine the weights of criteria. In suggested method a preference table is used to calculate the higher weights and lower weights of criteria. The higher and lower weights are used to calculate the fuzzy weightings and their precisions. Then, the final weights of criteria are calculated based on group decision making procedure using the fuzzy weights and their precisions. The significant novelty of introduced method is considering the precision of each expert who fill the pairwise tables two times by different questionnaires. The effect of people with higher accuracy will be considered with a higher impact in aggregation the results of individuals.