Representation and Evaluation of Influence Diagrams in a Common Framework for Interval Decision Analysis

This paper present a method for inclusion of influence diagrams within a common framework for analysing decisions under risk supporting interval-valued user statements. The method of inclusion support both modeling and evaluation, and the evaluation is performed through a conversion of influence diagrams into decision tree frames holding symmetric decision trees. The qualitative and explicit modeling of probabilistic independence cause implicit comparative constraints between variables, constraints which must not be violated in order to obtain correct evaluation results.