Evaluation of Wave-induced Liquefaction in a Porous Seabed: Using an Artificial Neural Network and a Genetic Algorithm -based model

Abstract

The evaluation of wave-induced liquefaction is one of the key factors for analysing seabed characteristics and the design of marine structures. Numerous investigations of wave-induced liquefaction have been proposed. However, most previous research has focused on complicated mathematical theories and laboratory work. In this study, we contribute an alternative approach for the prediction of the waveinduced liquefaction using an Artificial Neural Network (ANN) and a Genetic Algorithm (GA)-based model. Combined ANN and GA-based models are still a newly developed area in coastal engineering. In this study, a Genetic Algorithm-based approach is proposed to find optimal weights for the ANN model. It reduces the training time, and improves the forecasting accuracy for wave-induced maximum liquefaction depth, compared to using the normal ANN training procedure. Simulation results demonstrate the capacity of the proposed ANN model for the prediction of wave induced maximum liquefaction depth in addition to the proposal of GAs for training the ANN model.