Value of information and value of decisions

Abstract

This paper introduces an extended formulation of decision analyses, which provides an enhanced basis for the definition of the value of information, the value of actions and the value of actions and information analyses. The formulation of decision analyses is (1) extended by introducing an action implementation uncertainty and - following the reasoning of Raiffa and Schlaifer (1961) - (2) by considering both the information acquirement state and the action implementation state jointly and separately for the definition of the types of decision analyses. Decision value analyses are derived by explicitly distinguishing and addressing the cause of the expected utility gain namely by information, by actions or by both action and information. It is shown how different optimal sets of information and actions and their acquirement or implementations states, respectively, lead to different decision value classifications. Published studies and applied decision and decision value analyses are analyzed showing a diversity beyond the original definitions by Raiffa and Schlaifer (1961), however, also being less diverse in comparison to the introduced classification in this paper.