Spatial Modeling of TDS in Aquifer Using Genetic Algorithm Based on Artificial Neural Network Model (Case Study: Qazvin Plane Aquifer)

Recently, combination of the expert models such as genetic algorithm (GA) and Artificial neural networks models (ANNs) due to their capability in modeling of complicated and nonlinear phenomena specially the aquifer characteristics have attracted the hydrogeologists. In this study the capability of the continuous GA (CGA) based on hree layer perceptron network (GA-ANN model) has been investigated in the interpolating the values of total dissolved solids (TDS) in the Qazvin plane aquifer. Input variables in this model consist of the X and Y coordinates of recorded samples in aquifer and output variable is TDS value in corresponded site. To indicate the superiority of the GA-ANN model in spatial modeling of TDS, the ordinary kriging method (OK) that is known the best linear unbiased estimator (BLUE) also has been used. The obtained results indicate that the GA-ANN model has more capability in modeling spatial distribution of TDS values in the aquifer related to the OK method.

Spatial Modeling of TDS in Aquifer Using Genetic Algorithm Based on Artificial Neural Network Model (Case Study: Qazvin Plane Aquifer) Keywords:

Spatial Modeling , TDS , Genetic Algorithm , Artificial Neural Networks , Kriging , Qazvin Aquifer.