Abstract
Data envelopment analysis (DEA) is an approach based on linear programming to assess the relative efficiency of peer decision-making units (DMUs). Typically, each DMU is free to choose the weights of the factors used in its evaluation. However, the evaluator's preferences may not warrant so much freedom. Several approaches have been proposed to allow the incorporation of managerial preferences in DEA, but few address the additive DEA model specifically. This paper presents additive DEA models that use multi-criteria decision analysis concepts to incorporate managerial preferences, and presents the corresponding preference elicitation protocols. The models developed allow the incorporation of preferences at different levels: on valuing performance improvements, on introducing weight restrictions, and on finding adequate targets. These were application-driven developments, resulting from discussing modelling options and preliminary results with the top-level management of a retail chain in the context of an assessment of stores’ performance, also described in this paper.
Acknowledgements
This work has been partly supported by the POCI 2010 FCT grant EGE/58371/2004 and FEDER/COMPETE FCT grant MIT/MCA/0066/2009. The authors wish to express their gratitude to the management team of the client organization for their interest in this study, for their degree of involvement, and for all that we have learned from them. The authors also wish to thank the anonymous referees for their constructive remarks and suggestions.