Hongxiang Ren

Minimizing casualties and property losses from maritime emergencies requires a timely and reliable search and rescue (SAR) logistics network. Unlike conventional emergency logistics, maritime incidents are not only stochastic but also involve demand locations that evolve over time under ocean environmental influences. To address these characteristics, we develop a two-stage dynamic location–allocation–scheduling (2S-DLAS) model for maritime SAR operations under uncertain demand and time-varying demand locations. The model minimizes total rescue cost by jointly optimizing rescue base locations, the allocation of multiple SAR resource types, and the coordinated routing and task assignment of SAR platforms. It explicitly captures the dynamic coupling between travel time and ocean conditions and incorporates navigational restrictions into accessibility constraints. To solve the model, we use a Sample Average Approximation (SAA) framework and design an Improved Adaptive Large Neighborhood Search (IALNS) algorithm that integrates cross-stage nested neighborhoods to improve solution quality and computational efficiency. Numerical experiments based on real SAR cases from the South China Sea demonstrate the effectiveness of the proposed model and algorithm. Sensitivity analysis further provides managerial insights on strategic SAR deployment and multi-platform coordination, supporting evidence-based planning in maritime emergency response systems.

The joint optimization of vessel traffic organization and tugboat scheduling plays a critical role in improving the efficiency of port seaside operations. However, most existing studies are limited to single-channel environments or assume homogeneous tugboat fleets, whereas real-world ports are typically characterized by heterogeneous tugboat resources and, in some cases, complex multi-channel structures. To address this gap, this study investigates an integrated optimization problem combining vessel traffic organization with heterogeneous tugboat scheduling in a port with a dual one-way channel configuration. To capture the strong spatiotemporal coupling inherent in such operations, a bi-objective mixed-integer linear programming model is developed with the objectives of minimizing total weighted vessel waiting time and total tugboat fuel consumption. A tailored non-dominated sorting genetic algorithm incorporating an archive-guided adaptive multi-neighborhood search mechanism is proposed to efficiently solve the model. Computational experiments based on a northern Chinese seaport demonstrate that the proposed algorithm outperforms the CPLEX solver and several benchmark heuristic algorithms. The resulting Pareto solutions explicitly reveal the trade-off between vessel service efficiency and tugboat energy consumption, from which managerial implications are derived to support the formulation of coordinated scheduling strategies.

Condition monitoring and fault diagnosis of marine machinery are essential for ensuring the safety and stability of offshore operations. Traditional approaches, which typically rely solely on either physical mechanisms or large volumes of data, often struggle to achieve both high accuracy and adaptability to component degradation. To address this issue, this paper proposes a novel digital twin framework based on deep multimodal information fusion (DMIF-DT) that integrates the complementary strengths of a physics-based model (PBM) and a data-driven model (DDM). An attention-enhanced multi-scale convolutional network is developed to extract features from sensor measurements and calibrated simulation data. A low-rank fusion strategy is then employed to map the two modalities into a unified high-dimensional space for effective integration between the PBM and the DDM. Furthermore, a performance-degradation adaptive correction module, based on sensor measurements and probabilistic distribution features, is used to map health factors and feed them back to the PBM for component-level correction, thereby supporting periodic model updates. Experimental results show that the proposed DMIF-DT framework significantly reduces parameter prediction errors and achieves an average fault-diagnosis accuracy of 98.02% under diverse operating conditions. By uniquely integrating a performance-degradation adaptive correction module with attention-enhanced multimodal fusion, the framework bridges physics-based and data-driven insights, providing a new approach for the intelligent health management of marine machinery.

Physically-based turbulence modeling significantly enhances the realism of maritime simulator scenarios. We propose a novel importance sampling vortex particle method to simulate turbulent details within the SPH framework. The core of our method is a Monte Carlo estimator that approximates the global vorticity of the fluid system, addressing the computational cost associated with the global iteration of vortex particles. By constructing a probability function for importance sampling of vortex particles, we effectively suppress high-frequency variance fluctuations and improve simulation accuracy. Experimental results demonstrate that our importance sampling vortex particle method effectively reduces vorticity field variance by about 27% compared to uniform sampling, thereby improving simulation accuracy and producing visually prominent turbulent motion in the fluids. This study provides a theoretical reference for realistic ocean scene simulation in maritime simulators.

The safe operation and intelligent optimization of ships heavily rely on the predictive maintenance of marine engine performance. However, existing methods for predicting engine performance face limitations due to the various characteristics of ship engines, which depend on the sailing state and environmental conditions. To address this, a novel deep learning framework, the parallel channel-temporal attention-based temporal convolutional network (PCTA-TCN), is proposed to predict marine engine performance across varying load ranges. The PCTA-TCN framework introduces channel attention and temporal attention mechanisms in two independent branches, adaptively adjusting the focus on important features and temporal dependencies in the measured data. Parallel TCN backbones are used to extract high-level feature representations from different sequences. Finally, the model generates predictions by performing a nonlinear mapping on the extracted features at multiple levels. Experimental validation using sensor-acquired data demonstrates that PCTA-TCN consistently achieves an R² value exceeding 0.956 across all datasets. Compared to the baseline TCN model, the proposed method improves prediction accuracy by 13.95%, 49.81%, 15.64%, and 19.73% under four different load conditions. Furthermore, PCTA-TCN outperforms other state-of-the-art methods, particularly under complex and variable operating conditions.

In this study, a simplified radar echo signal model suitable for radar simulators and a Radar Cross-Section (RCS) calculation model based on the Physical Optics (PO) method was developed. A comprehensive radar target ship echo generation algorithm was designed, and the omnidirectional radar RCS values of three typical ships were calculated. The simulation generates radar target ship echo images under varying incident angles (0–360°), detection distances (0–24 nautical miles), and three common target material properties. The simulation results, compared with those from existing radar simulators and real radar systems, show that the method proposed in this study, based on RCS values, generates highly realistic radar target ship echoes. It accurately simulates radar echoes under different target ship headings, distances, and material influences, fully meeting the technical requirements of the STCW international convention for radar simulators.