Dynamic Programming Approach to Optimization of Site Remediation

Abstract

Environmental restoration is a matter of national concern. Decades of abuse by industry, agriculture, and the military have caused devastating contamination of the earth, air, and water. The Department of Energy alone will spend hundreds of billions of dollars on containment and restoration. It is imperative that restoration costs are minimized. Every dollar spent on restoration is a dollar that will not go toward research, a dollar that will not go to upgrade our nation’s infrastructure. The work presented here uses cost as a decision variable in restoration projects. Contaminated sites frequently vary from one point to another in type and level of contamination. In addition, a single piece of property may contain several distinct contaminated areas, each of which has characteristics unlike any of the other areas. Thus one should look at optimizing the selection of remediation technologies to address the variation. A methodology has been developed that will optimize the selection of remediation technologies based on cost. This methodology uses geostatistics and dynamic programming to break a site into discrete cells and then select the optimal sequence of remediation technologies.