Abstract

The Capacitated Warehouse Location consists of determining the number and locations of capacitated warehouses on a set of potential sites such that demands of predefined customers are met. Two typical assumptions in modeling this problem are: the capacity of warehouses is constant and that warehouses are able to truly satisfy customer demands. However, while these kinds of assumptions define a well structured problem from the mathematical modeling perspective, they are not realistic. In this thesis we relaxed such constraints based on the fact that warehouses can be built in various sizes and also warehouses can put in orders for unsatisfied customers' demand directly to the manufacturing plant with additional costs. This flexibility can lead to best decision making ability for managers and supply chain specialists to decide between higher capacity level with higher fixed and variable costs at the warehouse or direct ordering from the manufacturing plant. A new non linear integer programming formulation with staircase costs for multiple commodities in supply chain network is presented, and new method for linearizing the model is described. Computational results indicate that reasonably good solution can be obtained by the proposed linear model. Also for solving larger problems we developed a Tabu Search algorithm. The comparisons of the result between nonlinear/linear model and the Tabu Search algorithm are also presented.