Second Order Decision Analysis

摘要

I will discuss some aspects of risk analysis and decision making in practice. In such, the probabilities and utilities of consequences are often vague and imprecise. Furthermore, the support methods must allow for decision modelling with respect to different preferences and views, which further can be adjusted in an interactive fashion when considering calculated decision outcomes. There are several models for handling the natural impreciseseness in such situations, e.g., based on intervals probabilities or more elaborated approaches. I will discuss how second-order calculations add information, in a computational meaningful way, when handling imprecise representations, as is the case of decision trees or probabilistic networks. If time permits, I will also discuss some real life cases, when such methods significantly would have enhanced the analyses and why.