Decision Support with Partially Identified Parameters

摘要

In this paper, we investigate problem of determining a perference structure on the set of alternatives for a general class of single-stage, choice making models with imprecisely known parameters. A variety of decision making problems under certainty and under uncertaintly are modeled by the general problem formulation. The imprecisely known parameters can be, for example, attribute trade-off weights, value scores, value scores, probabilities, and utility values. Parameter imprecision is described by assuming that certain groups of parameters are members of given sets. This description forms the basis for a general and behaviorally relevant assessment model. Solution procedures for four important special cases of the general problem formulation are determined. A hypothetical automobile purchasing problem is used illustrate the decision aiding applicability of the results. (Author)