Kaisa Simola, Urho Pulkkinen, Jan-Erik Holmberg, Tony Rosqvist
Abstract:
Along with the adoption of risk informed decision making principles, the need for
formal probabilistic decision rule or criteria has been risen. However, there are
many practical and theoretical problems in the application of probabilistic criteria.
One has to think what is the proper way to apply probabilistic rules together with
deterministic ones and how the criteria are weighted with respect to each other.
In this report, we approach the above questions from the decision theoretic point
of view. We give a short review of the most well known probabilistic criteria, and
discuss examples of their use. We present a decision analytic framework for
evaluating the criteria, and we analyse how the different criteria behave under
incompleteness or uncertainty of the PSA model. As the conclusion of our
analysis we give recommendations on the application of the criteria in different
decision situations.