Automatic repair of inconsistent pairwise weighting matrices in analytic hierarchy process

The public procurement process involves contractor selection, where AHP (Analytic Hierarchy Process) has been widely applied to reflect decision makers’ (DMs) preference and priority on the basis of multiple criteria. Yet, AHP was criticized that DMs often cannot provide strictly consistent comparisons in a pairwise weighting matrix (PWM). The present study proposes an automatic repair procedure for the inconsistent PWM. The procedure seeks a substitute PWM, which can pass the consistency test, and yet being as close as possible to the primitive PWM. The repair is performed by a particle swarm optimization (PSO) algorithm, whose control parameters are tuned using the Taguchi method. The proposed procedure, by automatically adjusting the PWM with the approval of DMs, can greatly facilitate the process of contractor selection and thus expedite the entire procurement function. Two practical cases are used to demonstrate the advantages of the proposed procedure, in comparison with previous work.