Exact solution method for multi-stakeholder freight transportation systems under uncertainty

Abstract

At a time where efficient freight logistics are crucial to global commerce, integrated multi-stakeholder freight transportation systems play a pivotal role in ensuring timely delivery and operational adaptability under uncertainty. This study focuses on the tactical planning of such a system, which processes time-sensitive requests from both carriers and shippers. It orchestrates operations spatially and temporally, combining loads from various shippers into unified transport units. Our approach utilizes a two-stage stochastic programming model that effectively captures the uncertainties inherent in demand. The model formulates an efficient service network that not only meets the immediate logistical demands but also adapts to fluctuating conditions by leveraging available service capacities. To solve this complex model, we develop and implement an exact Benders decomposition-based algorithm. Our solution methodology incorporates several advanced techniques including partial decomposition, cut-lifting for both optimality and feasibility cuts, and various preprocessing steps including variable fixing and the use of valid inequalities. Additionally, we implement acceleration techniques that capitalize on the repetitive nature of our algorithm to enhance efficiency. We design and generate test instances inspired by real-world freight logistics, capturing key operational constraints and varying demand uncertainty levels. These instances enable a systematic evaluation of our model under diverse settings. We then perform extensive computational experiments. Our solution methodology demonstrates superior performance compared to a commercial solver. We also explore the impact of varying service availability among other parameters and the benefits of using stochastic modeling over deterministic approaches. These experiments underscore our model’s capacity to improve operational efficacy and responsiveness when dealing with uncertainty, thereby providing significant insights for both practitioners and researchers involved in freight logistics.