Evaluating deep uncertainty tools for complex decision problems

Abstract

Decision-making methods have been developed to support planning under deep uncertainty, but their suitability for complex decision problems has not been systematically assessed. We define complex decision problems as those in which decision-makers must navigate multiple interacting uncertain conditions and evaluate how interventions perform across these conditions, including potential synergies, trade-offs, and unintended consequences. This paper proposes two structural criteria for addressing such problems and evaluates a representative set of Decision Making under Deep Uncertainty (DMDU) methods against them, using a stylised case to assess each method’s capabilities. The results extend the established taxonomy of DMDU methods to include complexity of the decision problem as an explicit evaluative dimension and add classifications for two additional methods. The updated taxonomy provides decision-makers with a clearer basis for selecting or adapting tools to address the increasingly complex and uncertain decision problems they face