Locks are important components of a waterway system. To improve the efficiency of inland waterway transport, it is important to ensure ships passing locks without having to spend unnecessary waiting times at lock entrances. Meanwhile, with the trends towards digitalized and smart waterways, it is also worth investigating how the information availability could contribute to optimizing lock operations and ship arrivals on inland waterways. Therefore, this paper proposes an optimization method to schedule ships’ arrivals and their placements in locks on inland waterways, based on a mixed-integer programming model, and solves the optimization problem with large neighborhood search based heuristics. The optimization objective is threefold: first, optimizing the arrival sequence of ships at the locks; second, maximize the utilization of each lockage operation; and third, reducing the overall time that each ship spends from entering the waterway area till leaving the last lock on the waterway. Simulations are carried out to evaluate the performance of the proposed method.

Safe navigation of liquefied natural gas (LNG) carriers is critical for maintaining efficiency and mitigating catastrophic accident risks in modern maritime logistics, with heightened risks and operational challenges under nighttime conditions, which constitute the focus of this study. This study develops a hybrid methodological framework by integrating decision-making trial and evaluation laboratory–interpretive structural modeling (DEMATEL–ISM) with Bayesian networks to systematically identify, structure, and quantify risk factors associated with nighttime LNG carrier operations. First, key risk elements, spanning human, vessel, environmental, and managerial dimensions, are extracted via expert surveys and accident case reviews. DEMATEL–ISM is then employed to establish a multi-level hierarchical model that elucidates the direct and indirect causal relationships among these factors. The findings indicate that diminished ambient illumination and restricted visibility are deep-level triggers that propagate through psychological stress, visual performance degradation, and operational errors, ultimately increasing accident likelihood. Subsequently, a Bayesian network model is constructed by mapping the hierarchical structure onto probabilistic dependencies, allowing forward and backward inferences that pinpoint critical risk propagation pathways. The quantitative results reveal that risk chains predominantly originate from environmental disturbances and escalate when coupled with insufficient training, suboptimal management procedures, and human fatigue. These insights highlight the necessity for targeted interventions, including enhanced port lighting designs to minimize glare, scientific scheduling to counteract circadian disruptions, and strengthened on-site supervision through refined Vessel Traffic Services management and emergency response protocols. Overall, this integrated analytical framework provides both theoretical underpinnings and practical guidelines for improving nighttime LNG carrier safety, offering a valuable reference for researchers and stakeholders in maritime transportation risk management.