Two DEA Based Ant Algorithms for CMS Problems
| Main Authors: | Hossein Ali Akbarpour, Fatemeh Dadkhah |
|---|---|
| Format: | Article |
| Bahasa: | eng |
| Terbitan: |
, 2012
|
| Online Access: |
https://zenodo.org/record/1332316 |
Daftar Isi:
- This paper considers a multi criteria cell formation problem in Cellular Manufacturing System (CMS). Minimizing the number of voids and exceptional elements in cells simultaneously are two proposed objective functions. This problem is an Np-hard problem according to the literature, and therefore, we can-t find the optimal solution by an exact method. In this paper we developed two ant algorithms, Ant Colony Optimization (ACO) and Max-Min Ant System (MMAS), based on Data Envelopment Analysis (DEA). Both of them try to find the efficient solutions based on efficiency concept in DEA. Each artificial ant is considered as a Decision Making Unit (DMU). For each DMU we considered two inputs, the values of objective functions, and one output, the value of one for all of them. In order to evaluate performance of proposed methods we provided an experimental design with some empirical problem in three different sizes, small, medium and large. We defined three different criteria that show which algorithm has the best performance.