A normative (1) approach to decision making based on expected utility theory, some versions also incorporating Bayesian inference. It starts from the assumption that, for any pair of alternatives, the decision maker can express a preference, and that the preferences satisfy the following axioms, which accord with our intuitions about preferences. (1) Completeness: for every pair of alternatives x and y, either x is preferred to y, y is preferred to x, or the decision maker is indifferent between x and y; (2) Reflexivity: every alternative x is at least as preferable as itself; (3) Transitivity: if x is at least as preferable as y, and y is at least as preferable as z, then x is at least as preferable as z. The theory then requires a decision maker always to choose a maximally preferable alternative (which may not be unique). Many writers use the term synonymously with rational choice theory. Applied decision theory, taking into account the decision makers' utility functions, is called decision analysis. See also expected utility theory, intransitive preferences, money pump. Compare behavioural decision theory, psychological decision theory. DT abbrev.
(1) Completeness: for every pair of alternatives x and y, either x is preferred to y, y is preferred to x, or the decision maker is indifferent between x and y; (2) Reflexivity: every alternative x is at least as preferable as itself; (3) Transitivity: if x is at least as preferable as y, and y is at least as preferable as z, then x is at least as preferable as z.
