Hub location proplems under polyhedral demand uncertainty

Abstract

Hubs are points of consolidation and transshipment in many-to-many distribution systems that bene t from economies of scale. In hub location problems, the aim is to locate hub facilities such that each pairwise demand is satis ed and the total cost is minimized. The problem usually arises in the strategic planning phase prior to observing actual demand values. Hence incorporating robustness into hub location decisions under data uncertainty is crucial for achieving a reliable hub network design. In this thesis, we study hub location problems under polyhedral demand uncertainty. We consider uncapacitated multiple allocation p-hub median problem under hose and hybrid demand uncertainty and capacitated multiple allocation hub location problem under hose demand uncertainty. We propose mixed integer linear programming formulations and devise several exact solution algorithms based on Benders decomposition in order to solve large-scale problem instances. Computational experiments are performed on instances of three benchmark data sets from the literature.