A new approach for supplier selection in the presence of imprecise data: DEA with double frontiers

Supplier selection is a complex but important decision that requires careful review of various performance criteria. Traditionally, models of supplier selection have been based on cardinal data, with less emphasis on ordinal data. However, with the widespread use of production methods, such as the just-in-time method, recently more emphasis is placed on considering imprecise data—i.e., mixtures of interval and ordinal data. This article proposes to use data envelopment analysis (DEA) with double frontiers for supplier selection, a methodology which considers not only the optimistic efficiency but also the pessimistic efficiency of each supplier. Compared with the traditional DEA, the DEA approach with double frontiers can identify the best supplier accurately and easily without having to impose any weight limitations. A numerical example will be examined using the DEA approach with double frontiers to illustrate its simplicity and convenience for supplier selection and justification.