A practical method for improving consistency of judgement matrix in the Analytic Hierarchy Process

In this paper, an approach to improving consistency of judgement matrix in the Analytic Hierarchy Process (AHP) is presented, which utilizes the eigenvector to revise a pair of entries of judgement matrix each time. By using this method, any judgement matrix with a large C.R. can be modified to a matrix which can both tally with the consistency requirement and reserve the most information that the original matrix contains. An algorithm to derive a judgement matrix with acceptable consistency (i.e., C.R. < 0.1) and two criteria of evaluating effectiveness are also given.