ADVANCED MACHINE LEARNING FOR SHIP RESISTANCE PREDICTION

Ship resistance strongly influences fuel consumption and greenhouse gas emissions Traditional semi empiricalformulas (e g ITTC/Holtrop simplify complex physics and may miss nonlinear interactions across diverse hull shapesMachine learning can model nonlinear relationships, but purely data driven models often lack physical interpretabilityand may not generalize across different hull families Our approach combines physics with machine learning wepredict resistance components (e g frictional and wave and recombine them using physics based methods to maintaininterpretability and constraints