Abstract:

Aiming at the multi-attribute large group decision problems that attribute weights are completely unknown and attribute variables are continuous stochastic variables, a large group decision method is proposed. This method extends the 3σ principle of normal distribution to an arbitrary distribution and transforms stochastic attribute values into interval numbers. By using the fuzzy clustering algorithm in the interval number range, the large group is divided into several clusters. Then, the interval evaluation matrix of each cluster as well as the whole large group is defined and calculated. Ranking of alternatives is performed by using an uncertainty ordered weighted averaging operator. Finally, an example is used to illustrate the method.