Measuring the efficiency of network structures: Link control approach

• Published in: Computers & industrial engineering Vol. 128; pp. 437 - 446
• Main Authors: Hassanzadeh, Amir, Mostafaee, Amin
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.02.2019
• Subjects: (Non) Cooperative system; Link control; Network Data Envelopment Analysis (NDEA); Pull and push based system; Supply chain; Vendor managed inventory (VMI); Supply chain; Network Data Envelopment Analysis (NDEA); (Non) Cooperative system; Pull and push based system; Vendor managed inventory (VMI); Link control
• ISSN: 0360-8352; 1879-0550
• DOI: 10.1016/j.cie.2018.12.050

•This paper proposes an approach to compute the target intermediate products in Network DEA.•The overall and stage efficiency scores are obtained based on Link control notion.•Some of the existing approaches in Network DEA are reviewed according to the different scenarios of Link control notion. In network structures, the traditional models of data envelopment analysis (DEA) only use the external inputs and the final outputs in measuring the performance of the Decision Making Units (DMUs), and their internal operations are ignored. Therefore, these models cannot accurately determine the source of inefficiency inside the structures. To fix this problem, different Network DEA (NDEA) models have been proposed by scholars, yet they do not clarify which model is correct in a specific situation and why. This paper sheds light on this question. Furthermore, this paper examines the way intermediate products affect overall and stage efficiency scores of units in different scenarios. We identify six scenarios associated to the concept of Link control through different stages. Since different authors have used different approaches without clarification, these scenarios help us compute target intermediate products in NDEA. Four of the scenarios involve the assumption of “continuity between sending and receiving link value.” But this assumption is not made about two other scenarios, and hence the target values for these scenarios cannot be uniquely specified. Finally, a supply chain of soft drink producer has been used to show the reasonableness and applicability of the proposed method.