Multi-criteria Group Decision-making Method Based On Fuzzy Preference Relations, Research And Applications

The essence of multi-criteria decision-making is selecting the best alternative from lots of alternatives with different experts taking. First these experts provide their own opinions about the alternatives. Then their individual opinions will be aggregated by some approaches to selecting the best alternative. For the sake of the uncertain factors about the alternatives under different criteria this thesis investigates the multiperson-multicriteria decision-making based on fuzzy set theory and the linguistic evaluation of the experts.The fuzzy multicriteria group decision-making follows two steps before to achieve a final decision: aggregation and exploitation. How to define the fuzzy preference relation aad aggregation these relations to get the objective preference relation becomes the key to solve these decision-making problems. It needs lots of computations to get the fuzzy preference relation based on the existed approach and lack the analytical expression to compute the relation between the two fuzzy sets. Besides the measures that indicates the agreement between experts' opinions haven't presented. So through studying these above problems deeply we present a new weighed fuzzy preference relation and its computation method to get the individual preference relations. Then these individual relations are aggregated based the OWA operator to attain the objective preference relation. In addition to obtain the maximum degree of consensus or agreement between the set of experts on the solution set of alternatives some measures are presented.How to aggregate preference matrixes in linguistic, numerical, Interval-vahied form provided by experts is also studied in this thesis. Firstly, these preference matrixes in different form are transformed to the unified uncertain linguistic preference matrixes by use of the defined transformation function. Then, these unified uncertain linguistic preference matrixes are aggregated to a collective uncertain linguistic preference matrix based on the ULOWA (Uncertain linguistic Ordered Weighted Operator) proposed by us. The ranking order of alternatives is get on the conception of preference relation of uncertain linguistic variable.