AbstractAn optimization model is developed to locate pumping wells in a river bank filtration system to minimize the overall cost of pumping and treatment. The model combines the numerical solution of water flow and solute transport from the river to the pumping wells with a genetic-algorithm-based optimization technique. The water quality parameters considered in the study are the concentration of (1) suspended solids, (2) endosulfan, which is a widely used pesticide in agriculture, and (3) E. coli. It was observed that the fitness function increases at a faster rate near the river before attaining the optimum. However, the rate of decrease in the fitness function is gradual after the optimum. It is concluded that the genetic-algorithm-based optimization technique is an efficient and robust in the estimation of optimum well distance. Sensitivity analysis of the optimal distance and total cost due to variation in aquifer characteristics and decay rate is also presented.