Eliciting the utility of a decision maker is difficult. In this paper, we develop a flexible decision making framework, which uses the concept of utility robustness to address the problem of ambiguity and inconsistency in utility assessments. The ideas are developed by giving a probabilistic interpretation to utility and marginal utility functions. Boundary and additional conditions are used to describe a utility set that characterizes a decision maker's risk attitude. Reformulation and convergence results are given for the discrete and continuous specifications of the utility set. A portfolio investment decision problem is used to illustrate the basic ideas, and demonstrate the usefulness of the proposed decision making framework.
Department of Industrial Engineering and Management Sciences, Northwestern University, 2012