Groundwater remediation systems often cost millions of dollars at a single site. By replacing trial-and-error remediation design with a mathematical optimization algorithm, the most cost-effective policy may be determined, resulting in potentially large cost savings. This work applies genetic algorithms (GAs) to the optimal design of {\it in situ} bioremediation systems. GAs are search algorithms inspired by the idea of ``survival of the fittest.'' Because no derivative information is required, GAs can handle nonconvex, highly nonlinear, and complex problems. The optimization model combines the genetic algorithm approach with a two-dimensional flow and transport model for {\it in situ} bioremediation. In this work, a hypothetical {\it in situ} bioremediation design problem is developed for a homogeneous, isotropic aquifer. The objective of this design problem is to minimize the total cost of {\it in situ} bioremediation remediation (capital and operating costs), given the initial plume, by varying the injection rates at 17 potential wells subject to constraints on the final contaminant concentration at 73 observation wells and the hydraulic heads in the aquifer. At the end of the one-year remediation period, the contaminant concentration at the observation wells must meet the clean-up standard for the contaminant. In order to determine the impacts of time-varying designs on the total remediation cost, optimal designs are determined for both steady-state and dynamic cases. Additionally, the sensitivity of the optimal designs and costs on the water quality goals are evaluated. Results from both of these issues would allow the designer to choose effective and efficient remediation strategies.

Chan Hilton, A. B. and T. B. Culver. Optimal dynamic design of in situ groundwater bioremediation using genetic algorithms. Eos, Supplement to AGU spring meeting, May 26-29, 1998. Boston, MA, 1998.