4. Conclusion
This research studied on dynamic artificial neural net-work training using genetic algorithm optimization. From the tables and figures of the obtained and presented re-sults, we can conclude that the new solution construc-tion employed in this research with ability of altering and checking different weights, biases, number of hidden lay-ers and number of nodes in each hidden layer has led to the superior results of error rate and structure of ANN. The improvement obtained using the present method is due to the optimization of the structure of neural network si-multaneously with the weights and biases of the model instead of optimizing weights or structure of neural net-work in isolate as employed in other models.