Dry bulk river–sea intermodal transport is a critical consideration when connecting inland waterways and oceanic shipping, yet its efficiency hinges on precise vessel arrival time matching. The challenge of vessel arrival time matching has been exacerbated by existing research gaps. Current studies often focus on single vessel types or static scenarios, lacking integrated optimization of dynamic coordination between sea-going and river vessels, and failing to unify time and cost objectives. To address this, we develop a multiobjective scheduling model incorporating real-time arrival data from the dry bulk river–sea intermodal information platform to minimize total port time and operational costs. A heuristic genetic algorithm with adaptive weight adjustment (λ) is designed, achieving convergence within 200 iterations and a solution time of 33 seconds. This algorithm is validated under balanced conditions (λ=0.5) and is shown to yield 108.53 hours of total port time and 278,165.2 yuan in operational costs. Sensitivity analysis reveals a significant tradeoff: λ is reduced from 0.9 to 0.1, leading to an increase in port time by 1.42% but a reduction in costs of 3.03%. This reflects an improved flexibility in cost optimization as a result of resource manipulability. In contrast, port time is constrained by physical limits, such as loading/unloading efficiency. The framework developed provides practical decisional support for ports, with higher λ values (0.7–0.9) enabling rapid turnover in congestion and lower values (0.1–0.3) prioritizing cost economy. Future work should extend this approach to stochastic environments and incorporate multistakeholder coordination using game theory approaches.

Conventional railway emergency logistics frameworks are typically characterized by transport capacity adjustments to prioritize emergency material transportation. However, this paradigm frequently results in extended emergency response times and substantial delays in conventional freight operations. To address these limitations, an entire-process optimization strategy encompassing the Emergency Recovery Phase (ERP) and Post-Emergency Recovery Phase (PERP) was formulated, accompanied by a two-stage multi-objective optimization model. Diverging from conventional frameworks that necessitate operation plan reconfiguration for emergency train deployment, the proposed strategy streamlined operation plan replanning in the ERP through formation and loading plan optimization, while concurrently incorporating transportation cost-effectiveness in the PERP into the holistic optimization framework. The ERP submodel was designed to ensure the balanced allocation of limited emergency materials while achieving significant reductions in emergency response time. Subsequently, the PERP submodel incorporated dual considerations of transportation cost-effectiveness for railway carriers and cargo owners, while mitigating delay losses in conventional freight operations. To resolve this multi-objective optimization model, the Adaptive Variable Neighborhood Non-dominated Sorting Genetic Algorithm-II (AVNNSGA-II) was developed. The following results were obtained by this empirical study. (1) The ERP submodel attained emergency material satisfaction rates exceeding 51.28% across multiple disaster-affected areas while achieving emergency response time reductions of 6.16–19.22% relative to conventional railway emergency logistics frameworks. Notably, it demonstrated superior performance relative to road-based emergency logistics under different speed scenarios, with 55.9–69.4% response time reductions. (2) The PERP submodel effectively reduced delay losses in non-emergency freight operations by 50.49% through the implementation of differentiated transport prioritization mechanisms. (3) The superiority of this algorithm was confirmed with 97% of Pareto front solutions of AVNNSGA-II exceeding those of conventional NSGA-II. In conclusion, the proposed strategy is demonstrated to synergistically balance emergency response efficiency and transportation cost-effectiveness, thereby significantly enhancing railway emergency logistics performance. Furthermore, the integration of AVNNSGA-II with the multi-objective optimization model provides innovative perspectives for addressing large-scale rail freight allocation and scheduling challenges.