On the development of tools for decision analyses

Decision and Value of Information analyses can be highly computational demanding due to the exponential growth of the decision tree in dependency of the state and decision variables. Due to the advancement of computational methods in the last decades, decision problems with moderate complexity maybe solved by a software tool whereas highly complex analyses require tailored approaches, algorithms and large computational resources.

This paper contains designs of Value of Information analyses tools and its progression in complexity and approaches to cope with computational challenges. The design of the tool is focused on the basic elements of a decision analysis (namely the probabilistic information precision models together with action and the relevant system state performance models) and the description of the principal concepts for establishing the probabilistic models.

User interfaces with the generic probabilistic models and a software tool for moderately complex problems and for user defined models with a higher complexity are derived and introduced. With several examples, the utilization of the tool in different computational environments and the full exploitation of decision analyses results including the derivation of decision rules are demonstrated.