Optimal remediation with well locations and pumping rates selected as continuous decision variables

The design of a pump-and-treat groundwater remediation system can be solved as an optimization problem. A common approach in this formulation is to minimize the total cost of the pump-and-treat system, while defining the locations and extraction or injection rates of the candidate pumping wells as continuous decision variables. With this choice, the degree of freedom added to the optimization problem yields significant improvements on the solution. In this approach coupled solution of groundwater simulation models and optimization algorithms are required. The repeated use of the groundwater simulation models throughout the optimization cycle tends to be numerically complex and computationally costly when the governing equations are nonlinear. To overcome this drawback, we propose a new computational procedure, identified as progressive genetic algorithm (PGA), to solve the optimal design problem. PGA is a subdomain method, which combines standard genetic algorithm with ground water simulation models in an iterative solution process and provides a powerful tool for the solution of highly nonlinear optimization problems. Numerical examples are included to demonstrate the feasibility and efficiency of the proposed algorithm. Applications indicate that the proposed approach provides a feasible alternative for the solution of nonlinear optimization problems in groundwater management. (C) 1999 Elsevier Science B.V. All rights reserved.