Modeling and analysis of issues in hub location problem

Abstract

The hub location problem has been around for more than 10 years. The first mathematical model was formulated by O’Kelly (1986) which is a quadratic integer program. Since then, nearly all of the researchers in this area have concentrated on developing ’good’ linearizations. However, there are many aspects of the problem that need to be analyzed. In this dissertation, we investigate some of these issues. We first study the application areas of the hub location problem and clarify the underlying assumptions of the real world problems which lead to the customarily defined hub location problem. We identify a certain problem characteristic of cargo delivery systems, which is one of the major application areas of the hub location problem, which is not satisfactorily modeled by means of the customarily defined hub location models. We propose a new hub location model that captures the specific requirements that are particular to cargo delivery systems. Another issue that we concentrate on is the identification, modeling and analysis of the hub location problem under different performance measures, namely minimax and covering criteria. We propose new integer programming models for the hub location problem under minimax and covering objectives. Both of the new models are the result of a different way of approaching the problem and their computational performance is far more superior than the performance of the various linearizations of the basic models proposed for these problems in the literature.