back propagation neural network

The deformation modulus of rock mass (Em) is the most representative parameter of the pre-failure mechanical behavior of the rock material and of the rock mass .Due to the high cost and measurement difficulties of in situ tests, the predictive models using regression based statistical methods, back propagation neural networks (BPN) and fuzzy systems are recently employed for the indirect estimation of the modulus .Among these methods, the BPN has been reported to be very useful in modeling the rock material behavior, such as Em, by many researchers .Despite its extensive applications, design and structural optimization of BPN are still done via a time-consuming reiterative trial-and-error approach . However, in this research, the genetic algorithm (GA) is utilized to find the optimal parameters of BPN, such as the optimal number of neurons in hidden layer, learning rates and momentum coefficients of hidden and output layers of network. Then, the result is compared with that of trial-and-error procedure . For the purpose, a data base including118 data sets was employed from six dam sites locations in Zagros Mountains of Iran. According to the results, the GA -ANN model has higher accuracy than the trial-and-error model in the estimation of Em.