Abstract
Integer-valued data envelopment analysis (DEA) with alternative returns to scale technology has been introduced and developed recently by Kuosmanen and Kazemi Matin. The proportionality assumption of their introduced "natural augmentability" axiom in constant and nondecreasing returns to scale technologies makes it possible to achieve feasible decision-making units (DMUs) of arbitrary large size. In many real world applications it is not possible to achieve such production plans since some of the input and output variables are bounded above. In this paper, we extend the axiomatic foundation of integer-valuedDEAmodels for including bounded output variables. Some model variants are achieved by introducing a new axiom of "boundedness" over the selected output variables. A mixed integer linear programming (MILP) formulation is also introduced for computing efficiency scores in the associated production set. © 2011 The Authors. International Transactions in Operational Research © 2011 International Federation of Operational Research Societies.
Original language | English |
---|---|
Pages (from-to) | 741-749 |
Number of pages | 9 |
Journal | International Transactions in Operational Research |
Volume | 18 |
Issue number | 6 |
Early online date | 6 Sept 2011 |
DOIs | |
Publication status | Published - Nov 2011 |
Keywords
- data envelopment analysis
- integer-valued DEA
- efficiency
- mixed integer linear programming
- bounded outputs