Development of a Network Data Envelopment Analysis (NDEA) model for investigation of the internal structure of decision-making units

Data envelopment analysis has been in the center of attention due to independency of the production function. But the initial models of data envelopment analysis are incapable of examining the internal structure of the units and have a black-box view. One of the most common network structures is consecutive two-staged structure. Available models for evaluating this structure are mainly based on the decomposition approach, in other words, their priority is overall efficiency, and the efficiency of the stages is obtained by decomposing the total efficiency. In this paper, an attempt is made to develop a multivariate model that simultaneously considers the overall efficiency and efficiency of the stages. In addition, for multi-response mode, the models were developed to calculate the efficiencies and it was proved that in all models, efficiency scores range from zero to one, and a unit is efficient if only it is efficient in both stages. The presented models were used in an applied example and the results showed that the existing model performed more realistic evaluation than traditional models.