Multi-period hub network design from a dual perspective: An integrated approach considering congestion, demand uncertainty, and service quality optimization

Abstract

This study introduces a hub network design problem that considers three key factors: congestion, demand uncertainty, and multi-periodicity. Unlike classical models, which tend to address these factors separately, our model considers them simultaneously, providing a more realistic representation of hub network design challenges. Our model also incorporates service level considerations of network users, extending beyond the focus on transportation costs. Service quality is evaluated using two measures: travel time and the number of hubs visited during travel. Moreover, our model allows for adjustments in capacity levels and network structure throughout the planning horizon, adding a dynamic and realistic aspect to the problem setting. The inherent nonlinear nonconvex integer programming problem is reformulated into a mixed-integer second-order cone programming (SOCP) problem. To manage the model’s complexity, we propose an exact solution algorithm based on Benders decomposition, where the sub-problems are solved using a column generation technique. The efficacy of the solution approach is demonstrated through extensive computational experiments. Additionally, we discuss the benefits of each considered feature in terms of transportation costs and their impact on network structure, providing insights for the field.