Scheduled service network redesign in freight transportation: Joint demand and service selection under disruption

Natural disasters, including earthquakes, floods, hurricanes, and landslides, have profound impacts on transportation service networks, particularly disrupting rail and road transportation. Such events often damage critical infrastructure, such as terminals and their connecting bridges, tunnels, and railways, leading to prolonged service interruptions. To address these challenges, this paper introduces an innovative extension of the classical Scheduled Service Network Design (SSND) problem, termed the Scheduled Service Network Redesign (SSNR) problem. The SSNR is formulated as a mixed-integer programming (MIP) model that seeks to maximize total profit by balancing the revenue from selected demands against the costs of serving them under disrupted network conditions. The SSNR focuses on reconfiguring the post-disaster network by optimizing demand selection and assigning demands to available services under disruption conditions. A key distinction between SSND and SSNR is that they operate on two different physical network states for the definition of demands and services versus the execution of planning decisions. Although both demands and services are originally defined on the pre-disaster physical network, all subsequent decisions, including demand selection, service activation, and the construction of demand itineraries, must be carried out on the modified network after disruptions occur. This significantly increases the problem complexity, as some demands may become infeasible due to damaged terminals, rail links, or roads, while others may require rerouting through alternative itineraries that can lead to substantially higher transportation costs. To tackle these complexities, this paper proposes a Benders Decomposition (BD) reformulation to divide the SSNR into two subproblems. The efficiency of the BD reformulation is evaluated against an exact solver to demonstrate its performance. In addition, sensitivity analyses are conducted to assess the robustness of the SSNR model. The results provide actionable managerial insights for strengthening the resilience of transportation service networks in the face of natural disasters. These findings serve as a foundation for developing decision support systems, highlighting the effectiveness of the BD reformulation, and offering practical guidelines for improving scheduled service networks and ensuring continuity under disruptive events.