Sensitivity of Decisions to Probability Estimation Errors: A Reexamination

In the subjective expected utility model for decision making under uncertainty the decision maker often has difficulty assigning to the states probabilities with which he is completely satisfied. Six approaches that might be used to help the decision maker resolve his uncertainty about the values of his state probabilities are outlined. One approach tells the decision maker how much he must perturb his initial probability estimate in order to change his maximum expected utility alternative from the alternative originally best under the initial estimate. This approach is examined in greater detail and an iterative algorithm for solving the minimum-perturbation problem is presented.