Decision making using minimization of regret

AbstractWe are concerned with the problem of uncertain decision making. The paradigm of decision making using minimization of maximal regret (MMR) is introduced. We compare this technique with the classic Max–Min valuation method of decision making. We discuss a generalization of the MMR method leading to a parameterized family of minimal regret methods. We study this class in detail. An approach to decision making which combines valuation type decision functions with regret based decision functions is introduced. We apply the minimal regret method of decision making to situations in which our uncertainty profile is represented by a Dempster–Shafer belief structure