Estimation of comparisons based on interval AHP

In this paper, the proper judgment of a decision maker is estimated from the given comparisons based on Interval AHP. The given comparison cannot represent his/her judgment accurately, since s/he needs not to fix an exact weight to each item. Even if the weights are certain, s/he may not give an exact value to the comparison of two items by mistake. In the conventional Interval AHP, the former weight uncertainty is considered and the weights of the items is obtained as intervals. In order to consider the latter comparison uncertainty, this paper assumes that the weight of an item is constrained to be certain in the formulation of Interval AHP. The estimated comparison as the maximum interval ratio of the interval weights obtained by assuming that the weight of one of two corresponding items is certain is more reliable than the others. There are two reliable estimations of a comparison and the interval which is included in both of them is considered as the proper judgment.