Application of the FEM-ACO Hybrid Model in Determining the Optimum Locations for Pumping Wells

In this paper, an optimization simulation hybrid model is used to evaluate the optimum pumping policy in an artificial aquifer. In this study, the groundwater level drop was investigated in a hypothetical free aquifer with a surface area of 1.5 square kilometers and three different hydraulic conductivity with a thickness of 100 meters, as well as ten pumping wells. In order to estimate the groundwater level of the aquifer, the finite element method was used and to optimize the position of the wells, used the algorithm of the ant colony and finally the FEM-ACO model was presented. The position of wells with a specific discharge is optimized to minimize groundwater level losses in the aquifer. In this regard, the number of different number of ants, the effect of Foramen evaporation and the effect of the elite percentage of an ant collection on the value of the objective function were investigated. The results showed that sequencer antler algorithms and elite ant anchor algorithms with almost identical performance have the best performance among the antler algorithms, and after that, the antler community algorithm and the max-minimal ant antler algorithm are in the next ranks. All of the ant colony algorithms arrived too quickly into convergence, which makes this early convergence a suitable global optimum due to the use of chain constraints. Finally, after considering the proposed model, the proper position of the wells was determined. The results also showed that the maximum water loss in the aquifer is about 2.5 m.