Improved Real-Coded GA for Groundwater Bioremediation

Abstract

Cost-effective ways of remediating contaminated ground water by in situ bioremediation or other methods can be identified by coupling optimization and simulation methods. However, application of these methods to field-scale problems is limited by computational efficiency and by ease of use. In this paper, a more efficient genetic algorithm is developed and applied to in situ bioremediation of ground water. The algorithm involves a real-coded genetic algorithm (GA) coupled with two newly developed operators: directive recombination and screened replacement. This paper is the first application of a real-coded genetic algorithm (RGA) to ground-water remediation. The numerical results obtained for two bioremediation examples indicate that the directive recombination and screened replacement significantly improve the performance of RGA and that RGA performs much better than the standard binary-coded GA for the ground-water remediation problem. Because of the incorporation of interactions between the degrading microbes, oxygen, and contaminant concentrations, the equations for bioremediation are highly nonlinear. The RGA developed would also be expected to be more efficient for other highly nonlinear water resources problems.