Robust optimization
Fahimeh Baroughi; Soudabeh Seyyedi Ghomi
Abstract
In this paper the robust path centdian problem is investigated on tree networks with the same interval vertex weights for the both path center and path median problems. The used objective function in this paper is the simple sum of path median and path center problems. In the past research works the ...
Read More
In this paper the robust path centdian problem is investigated on tree networks with the same interval vertex weights for the both path center and path median problems. The used objective function in this paper is the simple sum of path median and path center problems. In the past research works the vertex weights for the both path median and path center location problems are disjoint. The used approach to compute the robust solution is the minmax regret criterion. In this method for any selected path on the tree, the maximum value of regret is minimized for all possible events of vertex weights. Using the minmax regret criterion, an algorithm with O(n^5) time complexity is presented to obtain a robust solution of the robust path centdian problem on tree networks. In this paper using the worst case scenarios for the path median and path center we obtain the worst case scenarios of robust centdian problem. Then we obtain a robust solution for this problem.
Location Modeling
Sepideh Taghikhani; Fahimeh Baroughi; Behrooz Alizadeh
Abstract
In this paper, the -product and t-state uncapacitated facility location problem is investigated. To be more precise, it is assume that each customer can request different products in a -state network. First, the mathematical formulation for the -product and t-state uncapacitated facility location ...
Read More
In this paper, the -product and t-state uncapacitated facility location problem is investigated. To be more precise, it is assume that each customer can request different products in a -state network. First, the mathematical formulation for the -product and t-state uncapacitated facility location problem with certain costs is proposed. Also, it is shown that this paoblem is NP-hard. Since in most real-world problems the input of data are often ambiguous and uncertain, we study the -product and -state uncapacitated facility location problem in which the facility set-up costs and customer service costs are fuzzy random variables. Using three criteria, probability-possibility, probability-necessity and probability-credibility, the -product and -state uncapacitated facility location problem is formulated as a quadratic programming. Finally, a practical example is given to illustrate the efficiency of the proposed approaches.