Guidance framework and software for understanding and achieving system robustness

Abstract

To aid decision making about environmental systems under deep uncertainty, robustness metrics are commonly used to represent system performance over a number of scenarios. However, there are many robustness metrics and many ways of generating scenarios, making it difficult to know which to choose in order to quantify system robustness and to make robust decisions. To address this shortcoming, we introduce a generic guidance framework to assist with the identification of the most robust decision alternatives, as well as the RAPID (Robustness Analysis Producing Intelligent Decisions) software package which is a consistent and easy-to-use implementation of the framework. We illustrate the framework and software package on a hypothetical lake pollution problem, known as The Lake Problem, showing how the framework and software package apply to several situations where decision-makers may or may not know which scenarios or robustness metrics to use.