Mathematical tools from the field of optimization have significant potential for reducing long-term monitoring costs and aiding site managers in making informed decisions on sampling strategies for sites undergoing long-term monitoring. A case study is presented that demonstrates the use of a Nondominated Sorted Genetic Algorithm (NSGA) for monitoring design at Hill Air Force Base (AFB). The method combines fate-and-transport simulation (although it can also be used only with historical data), plume interpolation, and adaptive search to identify the tradeoff between monitoring costs and mass estimation error. The method efficiently provides decision makers a direct representation of the tradeoff between monitoring objectives such as cost and error. Additionally, the most and least significant monitoring wells in a preexisting monitoring network are identified. Copyright ASCE 2004.