Multiobjective Optimization of Pump-and-Treat-Based Optimal Multilayer Aquifer Remediation Design with Flexible Remediation Time

Abstract

Groundwater flow model MODFLOW 2000 and contaminant transport model MT3DMS 5.0 have been directly linked with nondominated sorting genetic algorithm (NSGA) II using a novel technique to solve a multiobjective pump-and-treat groundwater remediation problem. In this technique, the system simulators in their compiled forms are directly coupled with the optimization algorithm. The performance of the model has been evaluated on a multiobjective groundwater remediation problem based on a hypothetical confined multilayered aquifer under steady-state flow condition. Two conflicting objectives, viz, minimization of remediation cost and minimization of remediation time subject to upper and lower bounds on extraction rates, upper bounds on hydraulic heads, and contaminant concentration levels, are incorporated into the optimization formulation. Incorporation of discrete cost components into the objective cost function renders it to be discontinuous and nondifferentiable, and the resulting Pareto front exhibits disjoints at a few locations. The disjointed feature of the Pareto front is used to study the impact of remediation time on remediation cost. The main contribution of this study is the finding of the significance of remediation time on optimal groundwater remediation design problems. The results suggest that remediation time plays a crucial role in the optimal aquifer remediation design. Unless compulsory, incorporation of remediation time as a decision variable in the optimization model may yield better results in minimizing the remediation cost. Another contribution is the linking of computer programs written in different languages. The linking technique used in this study has the potential of coupling any two or more codes written in different languages without writing interface programs.