Artificial neural networks and particle swarm optimization based model for the solution of groundwater management problem

Fulfilling the growing water demand, at domestic, industrial and agriculture level, is the most challenging task and groundwater plays the most important role for achieving this demand. In this scenario, proper management of groundwater resources is the most required act, as unmanaged groundwater extraction may cause shrinking of aquifer, sea water intrusion and water quality problems. The simulation-optimization approach is the most efficient way to solve any ground management problem where complex numerical models simulate the groundwater flow and/or contamination transport. The optimization model employs simulation for achieving the values of the groundwater head, velocity, concentration etc. This repeated use of the flow model increases the computational burden extensively and takes several hours to converge the final solution. In this study, Artificial Neural Network (ANN), Bagged Decision Trees (BDT) and Particle Swarm Optimization (PSO) models were developed and coupled for the management of groundwater resources. The Analytic Element Method (AEM) based flow model was developed and used to generate the dataset for the training and testing of the ANN model. These developed ANN-PSO & BDT-PSO models were applied to minimize the pumping cost of the wells. The results show that the ANN model can reduce the computational burden significantly and it is able to analyze different scenarios as well.