A novel approach for determining common weights in two division network DEA: a case study of Science and Technology Parks in Turkey

Abstract

Data Envelopment Analysis is a popular tool for assessing the relative efficiency of units, treated as a whole unit, without considering the internal structure of decision-making units. In contrast to the conventional models, Network DEA models consider the internal operations of units. In both models, the most favourable weights of the input-output factors for each unit are used. One of the main pitfalls of these models is that they require a set of n different vectors of these weights for the units, thus the factors may not have the same importance for each unit. To overcome this drawback and avoid ignoring some factors, several methods have been proposed to determine a common set of weights (CSWs) of the factors. However, none of the proposed approaches can determine the CSWs in network DEA models. This paper contributes to the existing literature by proposing a novel approach for determining the CSWs based on Entropy Method. Finally, the proposed approach is applied to demonstrate their applicability in both numerical and illustration examples. Moreover, a case of assessment of Science and Technology Parks in Turkey was used for the purpose of illustration.