Product-service supplier pre-evaluation with modified fuzzy ANP reducing decision information distortion

Abstract

The main challenge of new product-service supplier (PSS) selection is the lack of historical performance information of the suppliers. As one of the most popular techniques in supplier evaluation, the analytic network process (ANP) has an advantage in organising and analysing PSS pre-evaluation problems. However, the decision information distortion caused by matrix revision is a major factor affecting the wide application of the ANP when the consistency of the comparison matrix is unqualified. In this paper, a modified fuzzy ANP (F-ANP) with six criteria and 20 sub-criteria is suggested for PSS pre-evaluation. With the purpose of reducing decision information distortion in the pre-evaluation, the geometric scale is employed to improve the consistency of the judgment matrices and a linear approach is proposed for the unqualified judgment matrix revision. It is the first time that the decision information in the comparison matrix can be evaluated and retained as much as possible in generating and selecting the substitute using the proposed linear approach. The improved F-ANP model is verified in a real-life case study. Compared with the approaches in the literature, the linear approach shows advantages in retaining original decision information and the improved F-ANP outperforms the ANP and analytic hierarchy process models in terms of obtaining a stable rank of suppliers and distinguishing the importance of the criteria. 1 2015 Informa UK Limited, trading as Taylor & Francis Group.