Expected utility and case-based reasoning

This paper presents an equivalence result between expected utility theory (EU) and a modified version of case-based decision theory (CBDT). To be precise, it shows that a model constructed in EU can be embedded in a CBDT model, and vice versa. CBDT, proposed and axiomatized by Gilboa and Schmeidler (1995, Case-based decision theory, Quarterly Journal of Economics 110, 605-639), is related to case-based reasoning in psychology and artificial intelligence and considered as a descriptive theory of human behavior. In CBDT a decision maker remembers situations similar to the current problem and uses them to help solve it. This idea stems from bounded rationality and is similar in spirit to the satisficing theory of March and Simon (1958, Organizations, Wiley, New York) in the sense that the decision maker tends to satisfice rather than optimize. (C) 2000 Elsevier Science B.V. All rights reserved.