Handling Imprecise Information in Risk M anagement

Cost evaluation often constitutes a substantial part of the total risk analysis. Often, models that use decision theoretical methods at different levels in the risk evaluation process are unable to take into account situations where the available information concerning the consequences of different incidents is vague or numerically imprecise. Based on a more general theory for decision analysis, a method for cost evaluation is suggested. It includes well-founded and computable procedures that enable a risk manager to work with interval statements and comparisons. The method is easy to implement in a computer system and does not require the use of numerically overprecise statements of probability and cost. The evaluation results in an interval that expresses the maximum and minimum expected cost with respect to the estimations of the risk manager. The interval can be further investigated with respect to the range of values consistent with the estimations. The method extends a risk evaluation process currently in use in Telia AB (formerly Swedish Telecom).