Range estimation priority: a method to sift out weak-weighted criteria in AHP

In this paper, it is shown priority vectors are full of deviations and uncertainties. This occurs due to the decision-makers' uncertainties and errors in comparisons as well as inconsistencies in judgment matrices in Analytical Hierarchy Process (AHP). The contribution to evaluation goals of the criteria with very weak weights may be submerged in such deviations and uncertainties, which makes those criteria meaningless and should be sifted out. The concept and method to sift such criteria out are proposed in this paper. This method allows decision-makers to identify and sift out weak-weighted criteria; this is especially helpful for decision-makers who want to build a set of criteria for a specific problem from a large set of criteria for general purposes. With downsizing in criteria numbers, decision-makers are expected to make fewer errors in judgment, and it' s easier for judgment matrices to reach consistency. Moreover, the performance information of each alternative on those weak-weighted criteria is no more necessary, which further improves the usability of AHP. It' s proved that this method is feasible, convenient and robust by the numerical case provided in this paper.