Department of Mathematics
Developing a groundwater remediation system subject to uncertainty
Karen L. Ricciardi
Mathematics Program, Division of Natural Sciences
Bard College
A cost-effective groundwater pump-and-treat remediation design algorithm has been developed that takes into account the uncertainty of the hydraulic conductivity of a given aquifer. The resultant design is subject to both gradient constraints, which are linear with respect to changes in pumping rates, as well as concentration constraints, which are nonlinear with respect to changes in pumping rates. The uncertainty in the hydraulic conductivity is taken in to account in this model by using a multi-scenario approach whereby different hydraulic-conductivity fields, obtained through a representative sampling technique, are analyzed simultaneously. The tunneling method is a novel method of solving global-optimization problems of this form. This method has been modified to efficiently solve this optimization problem. An application to a hypothetical problem demonstrates the efficacy of this approach.
All interested persons are welcome.