Robust optimization of water infrastructure planning under deep uncertainty using metamodels

Abstract

Water resources planning and design problems, such as the sequencing of water supply infrastructure, are often complicated by deep uncertainty, including changes in population dynamics and the impact of climate change. To handle such uncertainties, robustness can be used to assess system performance, but its calculation typically involves many scenarios and hence is computationally expensive. Consequently, robustness has usually not been included as a formal optimization objective, but is considered post-optimization. To address this shortcoming, an approach is developed that uses metamodels (surrogates of computationally expensive simulation models) to calculate robustness and other objectives. This enables robustness to be considered explicitly as an objective within a multi-objective optimization framework. The approach is demonstrated for a water-supply sources sequencing problem in Adelaide, South Australia. The results indicate the approach can identify optimal trade-offs between robustness, cost and environmental objectives, which would otherwise not have been possible using commonly available computational resources.