Data envelopment analysis model and benchmarking in a hierarchical structure with dependent parameters

Subject Areas : International Journal of Data Envelopment Analysis

Sajjad Kheyri ¹ , Seyed Esmaeil Najafi ² , Bijan Rahmani Parchikolaei ³

1 - Department of Industrial Engineering, Central Tehran Branch, Islamic Azad University, Tehran, Tehran
2 - Department of Industrial Engineering, Science and Research Branch, Islamic Azad University, Tehran, Tehran
3 - Department of Mathematics, Science and Research Branch, Islamic Azad University, Tehran, Iran

Received: 2022-03-04 Accepted : 2022-07-14 Published : 2022-08-01

Keywords: Hierarchical Structure, Data Envelopment Analysis, Dependent parameters, Benchmarking,

Abstract :

Data envelopment analysis is one of the best methods to evaluate the performance of decision-making units. This method is also used for benchmarking. benchmarking is a tool to evaluate organizational performance with a learning approach from others, it is also one of the practical methods in continuous improvement of the benchmarking method. The importance of benchmarking in all industries is clear. This paper considers the after-sales service network of an automobile company in Iran to evaluate the model. According to the structure of this network, a hierarchical structure is considered for benchmarking. In this paper, the purpose is to provide a model for benchmarking decision-making units with hierarchical structure and dependent parameters. In the real world, most decision-making units have a hierarchical structure and this structure needs more attention by researchers also dependent parameters can have a high impact on benchmarking. The proposed model for the after-sales service network of an automobile company in Iran was implemented and the results show the high impact of dependent variables on benchmarking and has increased modeling accuracy. The accuracy of benchmarking is very important for the success of decision-making units and the results show that paying attention to the relationships between the parameters increases the accuracy of benchmarking and according to the proposed model, more accurate benchmarking can be achieved.