A comprehensive method for the centralized resource allocation in DEA. (January 2019)

• Record Type: Journal Article
• Authors: Sadeghi, J.; Dehnokhalaji, A.
• Is Part Of: Computers & industrial engineering. Volume 127(2019)
• Page Start: 344
• Page End: 352
• Publication Date: 2019-01
• Subjects: Data envelopment analysis -- Resource allocation -- Cost function -- Revenue function; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2018.10.011
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: Two generalized models are presented for the centralized resource allocation. An existing resource allocation model is extended to (re)allocate some (current) additional or less resources among DMUs. The total revenue and cost are optimized and in the absence of market prices, it leads to strongly efficient targets. Two empirical examples are presented to illustrate both formulated models. The proposed method is compared to several existing methods. Abstract: In this paper, we propose two new centralized resource allocation methods extending Lozano and Villa's method (2004) to more general cases. The main assumption in this study is that all Decision Making Units (DMUs) are operating under the supervision of a central DMU in order to introduce the input and output targets for all units in the next production period. This paper considers two planning ideas. The first one is maximizing outputs produced with future planned resources and removing all input inefficiencies as much as possible while making all units strongly efficient. The second one is optimizing total revenue and cost functions to reach the best performance of the system. All units are assumed to be able to adjust their input consumptions and output productions in the current Production Possibility Set (PPS). Finally, we show that the two proposed methods provide the same results when input and output prices are replaced with their corresponding shadow prices. The proposed methods are illustrated with two empirical data sets.