Two-level Efficiency Evaluation of Centralized Star Networks

چکیده مقاله

Conventional Data Envelopment Analysis (DEA) considers decision making units (DMUs) as black-boxes and ignores their internal structures in efficiency analysis. Network DEA includes the interrelations among a DMU’s components in its efficiency assessment, and evaluates the contribution of each component to overall network performance. In this paper, the star network structure, which is a common structure in real world applications, is introduced and the Network DEA methodology is specialized for assessing these structures. Assuming centralized management in the network, the central component has the upper hand, and the efficiency assessment is performed by giving higher priority to it. In the first level, the performance of the central component is evaluated by ignoring the sub-components. The efficiency of the whole network is evaluated in the second level so that the efficiency score of the central component is maintained. Efficiency decomposition and frontier projection are described using multiplicative and envelopment forms of the proposed model. The theoretical features and practical merit of the model are justified by mathematical discussions and a numerical example.