ABSTRACT

The U.S. EPA has estimated that remediation of contaminated soil and groundwater will cost on the order of tens of billions of dollars. Application of optimization has the potential to provide more efficient engineering solutions for remediation systems. Although much has been accomplished in this area in the last two decades, more work remains to solve the most challenging applications. This problem involves typical challenges associated with optimization of engineering systems, such as nonlinearity, nonconvexity, and multiple minima. The optimization problem is "simulator-heavy", that is, state variables are provided by computationally-intensive simulators. Thus, it is critical that efficient and robust solution methods be developed and brought to bear on the problem. Furthermore, uncertainty in the conceptual models, parameters, and initial and boundary conditions that form the basis for the simulators must be considered when solving the optimization problem. Decision analysis also may need to be incorporated into the solution framework, since conflicts among stakeholders can result in objective functions and constraints that are difficult to define precisely.

We will present the engineering and mathematical outline of the subsurface remediation design problem. We will discuss state of the art approaches for meeting some the challenges, including new optimization solution approaches, application of stochastic optimization, and incorporation of decision analysis into the problem. We will also discuss the development of test problems to be attacked by the engineering and mathematics community, as a means for benchmarking and comparing optimization approaches.