Multi-stage decisions under deep uncertainty

December 3, 2018
Versione stampabile

Venue: Polo scientifico-tecnologico Fabio Ferrari, via Sommarive 9 (Trento) - Girasole Room
Time: 14:00-16:00

  • Tommi Ekholm, Department of Mathematics and Systems Analysis - Aalto University, Finland

Abstract
Decision-making is particularly difficult if the decisions outcomes’ probabilities can be estimated reliably. As an example, one might have multiple experts or lines of evidence providing conflicting probability estimates that cannot be reconciled, leading to ambiguity in probabilities. This type of problem is one class of deep uncertainty. 
The presentation considers such multiple beliefs in a two-staged setting of decision-making. We present that the optimal decisions should be non-dominated across beliefs: one cannot improve the outcome predicted by one belief without making it worse in terms of another belief. The presentation discusses the dynamic consistency of non-dominated strategies and illustrates the concept with examples concerning medical research and climate change.

Biosketch
Tommi Ekholm holds a PhD in operations research and works as a Postdoctoral Research Fellow at Aalto University. His research topics cover decision-making under uncertainty, climate change and energy economics, sustainability and forest economics. In addition to his academic research, he has experience in real-world decision-support for policymakers on national and EU climate policy and in United Nations’ climate negotiations.