Predicting mooring system response using LDA-HBA-LSTM model

Accurate prediction of the response of mooring systems is essential to ensure the safety and reliability of offshore floating structures. Long Short-Term Memory (LSTM) networks have achieved promising results in time-series prediction tasks because of their excellent network architecture, but their performance is highly dependent on appropriate hyperparameter configurations. To address this challenge, this paper proposes a Latin hypercube and density-assisted Honey Badger Algorithm Optimizer (LDA-HBA), which increases diversity by utilizing Latin hypercube sampling for population initialization, introduces density to enhance the strength formulation, and adaptive crossover operator to further improve the exploration capability of the algorithm and prevent it from falling into local optima. Experimental results show that the prediction accuracy of LDA-HBA-LSTM is better than the traditional model and converges faster.