Decision making under severe uncertainty: models and algorithms

Research Group Systems

Location: CLT 701
Date/Time: 20th September 2011, 14:00 - 15:00

Speaker: Dr Matthias C.M. Troffaes, Department of Mathematical Sciences, Durham University. 

When dealing with severe uncertainty, often, it proves very difficult to quantify probabilities exactly, and to value the utility of rewards precisely. In this talk, I will provide a short introduction and overview of the field of imprecise probability, which aims to deal with such severe uncertainty, by means of a behavioural approach to probability, and relying on techniques from sensitivity analysis. I will then go on to discuss the problem of decision making under severe uncertainty. Some recent advances in solving sequential decision problems will be presented, particularly, techniques for backward induction, and foundational issues regarding the consistency of the solutions thus obtained.

Published: 30th August 2011