Group decision support by interval AHP with uncertainty-based hierarchical clustering

The group decision in this study follows the structure of AHP, where two kinds of criteria-comparisons and alternatives-comparisons are given by a member. Reflecting the difference of the comparisons, two models to aggregate individual decisions into a group decision are proposed and compared. The grouping process is achieved by hierarchical clustering, in which an individual is merged into the nearest cluster one by one. The similarity of individuals is measured by the uncertainty of the group decision, since it tends to be uncertain in case of different thinking individuals. The uncertainty is quantified by Interval AHP, which uses interval weights to reflect the uncertainty of the decision problem. Then, based on the increase of uncertainty by each step, the sub-groups are noticed. In order for individuals to recognize their standpoints and reconsider their judgments if necessary, the group decision in progress is open to them.