REDES NEURAIS ARTIFICIAIS E ALGORITMOS EVOLUCIONÁRIOS MULTI-POPULAÇÃO NA OTIMIZAÇÃO MULTI-OBJETIVO DA REMEDIAÇÃO DE ÁGUAS SUBTERRÂNEAS ARTIFICIAL NEURAL NETWORKS AND MULTI-POPULATION EVOLUTIONARY ALGORITHM FOR MULTI-OBJECTIVE OPTIMIZATION OF GROUNDWATER REMEDIATION

A groundwater remediation optimization problem through pump-and-treat technique is presented. Two objectives were pursued: 1) contaminant plume minimization; and 2) total cost minimization of remediation. Two multi-population evolutionary algorithms together with the artificial neural network (ANN) technology were em- ployed in order to solve this problem. Pump rates from wells were sent to ANN that computes the remained contam- inant mass left in the site. The ANN outcomes were sent to evolutionary algorithms to execute optimization process.