Application of novel hybrid artificial intelligence algorithms to groundwater simulation

Depletion of groundwater resources is nowadays being discussed as one of the most significant challenges in this field. Thus, simulation of the available groundwater resources seems necessary for taking reliable management measures for the plains. In the present study, a combination of various metaheuristic algorithms including feed-forward artificial neural network (ANN), the algorithm of innovative gunner (AIG), and black widow optimization (BWO) algorithm was employed to simulate the groundwater of the Selseleh plain located in the southwest of Iran during the period 2008-2018 on a monthly time scale. Furthermore, chicken swarm optimization (CSO) was adopted to optimize the weight coefficients. The data were divided with 70% being used for training purpose and the remaining 30% for test validation. Different statistical indices including the coefficient of determination (R-2), root-mean-square error (RMSE), mean absolute error (MAE), Nash-Sutcliffe efficiency (NSE) coefficient, and percentage of bias (PBIAS) were utilized for evaluation of the efficiency of modeling. The results illustrated that all the three hybrid models had better results in combinatory patterns. Moreover, the evaluation criteria suggested that among all the applied models, the ANN-AIG model was the only one possessing all statistical indices (e.g., R-2 = 0.964-0.995, RMSE = 0.273-0.71 m, MAE = 0.219-0.059 m, NSE = 0.886-0.978, and PBIAS = 0.002-0.001) in the validation stage. In a nutshell, this study suggested that among all the recommended hybrid models, ANN-AIG, ANN-BWO, and ANN-CSO reduced the simulation error compared to the standalone ANN model with (12.78-20.26)%, (10.13-17.61)%, and (9.27-1675)%, respectively.