| Registro completo |
| Provedor de dados: |
AgEcon
|
| País: |
United States
|
| Título: |
Polarity in DEA Models
|
| Autores: |
Agrell, Per J.
Bogetoft, Peter
Tind, Jorgen
|
| Data: |
2005-11-18
|
| Ano: |
2000
|
| Palavras-chave: |
Data Envelopment Analysis
Polarity
Research Methods/ Statistical Methods
|
| Resumo: |
This article discusses models in data envelopment analysis (DEA) relaxing the standard convexity assumptions. The basic model treats mutually incomparable pairs of sets to be generated by a procedure proposed earlier. Each pair consists of a consumption set and a production set of feasible input-output combinations. Two fundamental operations by the procedure are based on intersection and convex hull generation in the input-output space. A polarity analysis is performed which, subject to the usual assumptions about free disposability and nonnegativity, appears fruitful to do in the framework of blocking and antiblocking sets. It is shown how this leads to an interchange of the above operations extending some classical results from convex analysis. The last part of the paper presents a pair of linear programming models calculating a Farrell productivity index based on a preceeding application of the procedure. This is a generalization of the classical linear programming models in DEA subject to standard assumptions about convexity.
|
| Tipo: |
Working or Discussion Paper
|
| Idioma: |
Inglês
|
| Identificador: |
19151
http://purl.umn.edu/24188
|
| Editor: |
AgEcon Search
|
| Relação: |
Royal Veterinary and Agricultural University>Food and Resource Economic Institute>Unit of Economics Working papers
Unit of Economics Working Paper 2000/4
|
| Formato: |
20
application/pdf
|
