Design and Evaluation of Neural Networks for Pavement Rutting

摘要

In this study, 3-layer and 4-layer Neural Networks (NNs) are designed and evaluated for the rutting performance of asphalt pavements. To generate data for the neural network model, a total of 519 sets of processed data are obtained from mix design information and laboratory tests. The factors affecting rutting of asphalt concrete are used to define the domain of NNs. Networks are designed and trained using the Levenberg-Marquardt minimization algorithm. Using randomly generated weight factors to initialize the training algorithm, histograms are compiled and outputs estimated using a statistical approach. An excellent agreement is observed between test data and simulations. It is believed that the developed NN design procedure will be a useful tool in the study of pavement design and wear.