Risk based corrective action design using genetic algorithms

Considerable resources have been expended in attempting to restore sites with contaminated groundwater. In the past, the cleanup goals were often established without regard to risk, mandating remediation of groundwater to background or non-detection or maximum contaminant limits. These are often difficult or impossible to achieve and have made site restoration prohibitively expensive. In response to these concerns, risk-based corrective action (RBCA) is becoming a method of choice for remediating contaminated groundwater sites. Under RBCA, the risks to human health and the environment due to contamination are evaluated and measures taken only to minimize the risk to acceptable levels. A major difficulty in RBCA is negotiating an appropriate risk-based limit and a reasonable corrective action approach, particularly given all of the sources of uncertainty in predicting risk. To aid in this process, a new framework for negotiation is being developed that combines an optimization model with simulation models in order to develop risk-based remedial designs that are both cost effective and reliable. The model combines contaminant fate and transport simulation models and health risk assessment procedures with genetic algorithms to simultaneously predict risk and propose cost effective strategies for reducing the risk. To use the model, stakeholders first negotiate the objectives of the remediation, which may include minimizing risk, minimizing cost, and minimizing cleanup time. Then any constraints such as hydraulic head limits or social or economic constraints are considered. In this paper, the steps are demonstrated using a case study. Copyright ASCE 2004.