Information Matrix-Based Adaptive Sampling in Hull Form Optimisation

Hull form optimisation involves challenges such as large design spaces, numerous variables, and high nonlinearity. Therefore, that only use global approximate models alone cannot yield desirable results. An information matrix-based method is proposed for dynamically embedded local (IM-DEAM) in this paper, which uses the Gaussian-function matrix to extract one or more subspaces additional sampling a Latin hypercube (LHD) adaptive sampling. In addition, prevent overfitting by some spaces because of uneven distribution samples, are identified enable accurate description subspaces. The effectiveness robustness validated analysed applying optimise mathematical functions hull DTMB 5415. results demonstrate effective improving accuracies can produce reliable