Ramalan Cirian Reologi Campuran Berasfalt Menggunakan Rangkaian Saraf Tiruan

The primary objective of this study was to develop two types of artificial neural network models, namely: multilayer feed-forward neural network and radial basis function network to predict the rheological properties of asphalt mixtures in terms of i) complex modulus, E* and ii) phase angle, δ. This study also conducted to investigate the accuracy of two types of models in predicting the rheological properties of asphalt mixtures by means of statistical parameters such as the coefficient of determination (R²), mean absolute error (MAE), mean squared error (MSE) and root mean squared error (RMSE) for each developed models. The prediction models were developed using E* and δ data that was obtained from a previous study done by a group of researchers at the Nottingham Transportation Engineering Centre. Based on artificial neural networks analysis, both models show good correlations in predicting of rhelogical properties of asphalt mixtures with the R² values exceed than 0.99. A comparison between two types of artificial neural network reveals that radial basis function network is more accurate compared to the multilayer feed-forward neural network with higher of R² values and lower MAE, MSE and RMSE values.It was concluded that the artificial neural networks, which did not rely on mathematical expressions, can be used as an alternative method for predicting the rheological properties of asphalt mixtures.