A Robust Scenario Based Approach in an Uncertain Condition Applied to Location-Allocation Distribution Centers Problem

INTRODUCTION Facility location decisions are costly and difficult to change. One of the most important issues in facility location problem and specifically supply chain management is locating and allocating the distribution centers. These strategic decisions are critical factors of whether materials will flow efficiently through the chain network. The location of distribution centers are affected by parameters such as demands and transportation costs. Owing to the fact that each parameter of this problem can change sensitively in the period of time, so decisions related to the design can be very important and effective on whole supply chain network. The best optimal location of these nodes can surely save the transportation costs. Location-allocation (LA) problem in facility location problem is to locate a set of new facilities in a fashion that the total distance from facilities to customers and consequently the total cost is minimized. LA problem has been considered for many years because of its broadly realistic application. In real cases, we should ABSTRACT: The paper discusses the location-allocation model for logistic networks and distribution centers through considering uncertain parameters. In real-world cases, demands and transshipment costs change over the period of the time. This may lead to large cost deviation in total cost. Scenario based robust optimization approaches are proposed where occurrence probability of each scenario is not known. It is supposed that in this case there would be budget constraints and also holding the products in the distribution centers until sending them to the retailers’ destinations results additive cost that can be defined as inventory control cost in the model. In this paper, uncertainty is defined by different scenarios. Some robust approaches are presented that can be applied in location-allocation problem. The robust scenario based approaches like absolute robust and robust deviation are applied in locationallocation problem. Also a new robust approach is proposed that outperforms the existing classical approaches. The mean expected model has been discussed and compared to the robust proposed approaches. A numerical example illustrates the proposed model and the results have been reported. Finally the comparison of results shows the efficiency of proposed robust approach in comparison of classical approaches and also mean expected model.