نوع مقاله : پژوهشی
دانشکدهی مهندسی صنایع، دانشگاه یزد
عنوان مقاله [English]
In this article, authors propose a mixed integer bi-criterion linear programming model for facility locations-network design problem under uncertainty in the competitive environment. This model takes into consideration the possibility of competition and interruption in servicing availability. The objective functions are of cost minimization and facility attraction maximization. The cost components are facility construction cost, the routes' enhancement, improvement and or development expenses, and transportation costs. Due to the facts that customers' demands as well as transportation expenses are uncertain authors have employed robust type modeling of the problem taking scenario approach into consideration. We also have taken into considered the concept of reliable network design and attraction function for facility location in the competition environment. In order to show the application of the proposed model, a real case study discussing the facility location design and facility implementation for a new CT-Scan system in Fars province was studied. The result of this study indicates that Fars providence has capacity for four CT-Scans that can be positioned in the cities of Fasa, Marvedasht, Estahban and Shiraz. For sensitivity analysis purposes, two parameters are considered separately and simultaneously. Parameter can have impacts on the cost objective function and hence on the final solution of the problem. The aim of this research is obtaining an answer that is less sensitive to the changes in data and meantime increases the system's reliability. A tradeoff between the solution robustness and model robustness for various values of parameter is demonstrated by the figure in the body of the article. This tradeoff can help the decision makers in determining suitable weight for Due to the fact that
when the number of scenarios increases the use of exact methods of solutions are impossible for the reason of being an Np-hard problem one needs to employ a meta-heuristic approach to solve this problem. This can be considered as a future research for those who want to work on the extension of this problem.