Review on Applications of Neural Network in Coastal Engineering

Artificial Neural Networks (ANN) finds wide variety of application in solving problems related to coastal engineering. Its ability to learn highly complex interrelationship based on provided data sets with the help of a learning algorithm along with built in error tolerance and less amount of data requirement, makes it a powerful modeling tool in the research community. Large number of studies has been carried out in various fields like prediction of wave parameters, tidal level and storm surge, estimation of design parameters, liquefaction depth and scour depth to name a few. Various forecasting, estimation and supplement to the missing data studies carried out from different perspective ranging from, the sensitivity analysis to check the effect of input parameters and reduce the input size by discarding less effective ones; reducing the input size by using data assimilation techniques like principal component analysis to decrease the computational time requirement; usage of updated algorithms to overcome the problem of overfitting and overlearning, thereby increasing the network efficiency; has been carried out successfully, establishing ANN as an strong alternative to the data demanding and time consuming hydrodynamic and numerical models. As the validity of ANN to the ocean engineering applications became increasingly evident studies were incorporated in practical applications as well. Studies are being carried out to merge ANN with other AI techniques of Genetic Programming and Fuzzy Logic approaches to overcome the setbacks observed in ANN models. The studies have successfully shown that ANN can be applied to solve vast problems related to ocean engineering problems by meticulous selection of data, input parameters, network architecture and learning algorithms.