In this paper, a Genetic-Based Neural Network (GBNN) is employed to predict the soil-water characteristic curve of unsaturated soils. A three-layer network has been trained by genetic algorithm and its topology is determined by trial and error. The network has five input neurons, namely, initial void ratio, initial gravimetric water content, logarithm of suction normalized with respect to air pressure, clay fraction and silt content. The network has five neurons in the hidden layer and the only output neuron is the gravimetric water content corresponding to the assigned input suction. Results from pressure plate tests carried out on clay, silty clay, sandy loam and loam, compiled in SoilVision software, was adopted as a database for training and testing the network. For this purpose, and after data digitization, a computer program coded in visual basic was developed and used for the analysis. Finally, neural network simulations are compared with the experimental results, as well as models proposed by other investigators. The comparison indicates the superior performance of the proposed method for predicting the soil-water characteristic curve.
)