EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
2181. Path-based DEA models
Invited abstract in session WD-48: DEA methodological developments II, stream Data Envelopment Analysis and its Application.
Wednesday, 14:30-16:00Room: 60 (building: 324)
Authors (first author is the speaker)
1. | Jakub Hrdina
|
Faculty of Mathematics, Physics and Informatics, Comenius University Bratislava | |
2. | Maria Trnovska
|
Dpt. of applied mathematics and statistics, Faculty of Mathematics, Physics and Informatics, Comenius University in Bratislava | |
3. | Margareta Halicka
|
Comenius University in Bratislava |
Abstract
Data envelopment analysis (DEA) models appear in the envelopment and the multiplier form, which are in a primal-dual relationship. Both forms are essential for the proper interpretation and application of the model. New models are typically presented in the envelopment form, yet if they are not linear, it can be more challenging to acquire their dual counterparts. In this contribution, we focus on path-based DEA models, which search for the benchmarks by specifying various parametric paths running from the assessed unit to the boundary of the technology set. The class of path-based models includes well-known models such as the BCC input and output models, the hyperbolic distance function model and the general directional distance function model. For a general scheme of path-based models in the envelopment form we derive the general multiplier form of path-based models. Since path-based models in general are not linear, we study the duality relations between the primal-dual pair of envelopment and multiplier path-based models. Finally, we focus on the multiplier form of the BCC input and output models, the general directional distance function model and the general hyperbolic measure model using the derived general multiplier form of path-based models.
Keywords
- Data Envelopment Analysis
- Mathematical Programming
- Convex Optimization
Status: accepted
Back to the list of papers