Application of Artifitial Neural Network to Predict the Resilient Modulus of Stabilized Base Subjected to Wet-Dry Cycles

One of the most important input parameters for design of pavement structure based on mechanistic-empirical method is resilient modulus of different pavement materials. This research aims to provide a model for predicting resilient modulus of stabilized base subjected to wet-dry cycles based on Artificial Neural Network (ANN). Dataset for developing ANN consists of a total of 704 records and five attributes which is adopted from Maalouf et, al. (2012). A Feed-Forward back propagation neural network ws employed to predict resilient modulus of stabilized base with respect to five input parameters including the number of W-D cycles, the ratio of free lime to SAF (Silica, Alumina and Ferric Oxide compounds in the cementitious materials), the ratio of maximum dry density to the deviator stress. The results show that the artificial neural network can be used as a powerful and accurate tool to predict the resilient modulus of stabilized base in presence of wet-dry cycles. Also comparison of ANN with Support Vector Machin (SVM) confirms that the accuracy of ANN is superior to SVM method.