A hybrid multi-attribute decision-making and data envelopment analysis model with heterogeneous attributes: The case of sustainable development goals

This study presents an integrated multi-attribute decision-making (MADM) and data envelopment analysis (DEA) framework for solving problems with heterogeneous attributes. We classify the heterogeneous attributes into desirable and undesirable classes and provide a model for aggregating the attributes' weights and the alterna-tives' scores. The proposed model is initially designed as a Multiple Objective Decision Making (MODM) problem with a Data Envelopment Analysis (DEA) policy and then reformulated as a linear programming model tackled through a goal programming approach. We apply the proposed model to a set of European countries based on their fulfillment of the 17 Sustainable Development Goals (SDGs) defined by the United Nations. We show the proposed approach minimizes computational efforts and complexities and maximizes the participation and satisfaction of decision-makers. We compare the rankings derived from our model with those obtained from standard MADM techniques such as Euclid and TOPSIS. We illustrate how the different normalization methods are applied to condition the discrimination power of the models and analyze the reversals triggered by TOPSIS relative to the other techniques. We conclude by noting that our model does not rely on the weights defined by the experts to determine the ranking, which constitutes a significant advantage over the standard MADM techniques in strategic evaluation environments.