Synchromodal Bi-Objective Two-Stage Stochastic Optimization for Container and Power Unit Allocation

Abstract

Traditionally, logistics planning suffers from a ”siloed” approach where container repositioning andpower-unit deployment are managed in isolation, ignoring the tight practical coupling between these decisions. This project challenges that convention by proposing an integrated two-stage stochastic mixedinteger programming model for landside intermodal logistics. Our approach jointly allocates resources across both the drayage and trucking networks, explicitly accounting for demand uncertainty. The model is structured to provide stability and flexibility: Stage 1 establishes the baseline for repositioning, leasing, and initial deployment, creating a hard link between the networks through leasing decisions. Stage 2 allows for recourse adjustments once actual demand scenarios unfold. By treating the planning process as a bi-objective problem, we weigh drayage costs against trucking costs. Through the use of the ε-constraint method, we generate Pareto-efficient frontiers that clearly illustrate the superiority of a coordinated strategy. The results highlight that optimal total system costs are achieved through integration, standing in sharp contrast to the higher costs associated with isolated decision-making. Validation through a regional case study in Florida confirms that this coordinated planning approach significantly lowers expected system costs. Furthermore, it offers critical insights into operational bottlenecks, clarifying when certain constraints become restrictive. While the submodels are individually functional, the integrated execution yields strictly better performance across the entire network.