Sensitivity Analysis Of Network DEA And Its Application To The Weight Choice Of Internal Units

DEA (data envelopment analysis, DEA) has been applied as an objective mathematical programming method of efficiency evaluation in many economic activities when processing production system with multiple inputs and multiple outputs. In recent years, its frontier has gradually shifted from a concern of overall efficiency to the attention of DMU (decision-making unit) internal structure, i.e. the so-called network DEA research. Besides, one of the topics of interests in data envelopment analysis (DEA) is sensitivity and stability analysis of any specified Decision Making Unit (DMU). Based on the basic framework of the network DEA, sensitivity analysis combined with weight choice of sub-DMUs is studied in this article.In order to assess the sensitivity and stability of internal unit efficiency classification in network DEA, this paper presents a new method for measuring the sensitivity based on the network DEA framework. The core idea is that the efficient frontier could be moved without affecting the original efficiency classification. By using the proposed models, one can analyses the stability of intermediate product flow in a mixed network structure of sub-DMUs within a DMU. With the model, this paper allows calculation of the relative stability radius of the entire DMU and each sub-DMU, which can obtain an objective method for relative weight choice of the internal sub-DMUs in a mixed network DEA structure. Finally, an illustrative example is given to show the use of our models.