Ship Equipment Modeling Method Based on Neural Networks and Particle Swarm Optimization

To address the challenges in ship equipment control simulation and the complexity of building equipment mechanism models, a data-driven equipment modeling approach is proposed. This method aims to enhance model prediction accuracy and generalization capability through the construction of neural network models and optimization via Particle Swarm Optimization (PSO). First, a data-driven neural network model is developed using real ship operational data. Then, the connection weights of the neural network are optimized using PSO to reduce prediction errors and improve model performance. Experimental results demonstrate that the proposed method significantly improves the prediction accuracy of critical ship equipment, reduces the Mean Absolute Error (MAE), and increases the Coefficient of Determination ($\mathrm{R}^{\mathbf{2}}$). The proposed neural network and PSO-based modeling approach offers a novel perspective and effective tools for the simulation and optimization of ship equipment.