Estimating of the marshall stability values of modified bituminous mixtures by artificial neural network

Increasing traffic volumes along with the adverse effects of climatic conditions on roads in our country are causing permanent deformations in road coverings. In this study; Models were constructed using artificial neural networks and regression techniques to estimate the marshall stability values of modified bitumen for the extension of the life of the asphalt coating. The penetration, viscosity, softening point, bitumen weight in water and bitumen weight in weather values of the modified bitumen samples were used as independent variables when modeling. Independent variables were used in combinations of 2, 3, 4, and 5 among themselves to increase the number of prediction models. The generated models were analyzed by Matlab program and regression models and artificial neural network methods were compared. As a result of the analyzes made, it was observed that the models constructed with artificial neural network gave better results than the models constructed using regression techniques.