 Decision Management Solutions® |
||||||||||||
|
Decision Making under Risk and Uncertainty
|
At Robust Decisions we assume all information used in decision-making evaluations is uncertain. In fact, the algorithms that underlie the suite of Accord tools are based on Artificial Intelligence methods specifically designed for handling uncertainty. The Accord tool suite has been specifically designed to support decision making under risk and uncertainty. This page is a brief introduction to how uncertainty is managed through easy-to-use graphical interfaces. The difference between risk and uncertainty is discussed on an additional page. It makes no difference if an evaluation is based on a wild guess or on detailed simulations; it is the decision maker's belief in the alternative's ability to meet the criteria that is the basis for his decision. Team members project their knowledge, values, roles, and experience into an uncertain future and make decisions based on their beliefs in uncertain evaluation estimates. Uncertainty and belief are not bad, they just need to be recognized, supported, and managed. Most decision-making methods can't do this. A part of this philosophy is that uncertainty avoidance is not possible, but uncertainty management is. To help manage belief, divide evaluation into two independent parts: the assessments of Level of Criteria Satisfaction and Level of Certainty. You are used to thinking of evaluation as determining the Level of Criterion Satisfaction. This is what you evaluate in a Decision Matrix (Pugh's method) or when using any other method in which evaluation is a single, deterministic value. But a second dimension—Level of Certainty—is added to each evaluation to account for certainty. This can be shown on the two dimensions of a Belief Map.
In the example Belief Map assume the team is evaluating an alternative versus a qualitative criterion. In the example, Anne has represented her evaluation with a dot that shows she believes that alternative has a medium (M) to high (H) Level of Criterion Satisfaction. Based on her knowledge of the alternative her Level of Certainty about this evaluation is medium. Anne's evaluation is the large dot in the figure. For this qualitative criterion, think of certainty as equivalent to knowledge: Anne's self-assessed knowledge about the alternative is medium. John, on the other hand, bases his evaluation on his past work with similar systems, so he is more certain and has higher satisfaction. Lisa is not impressed, and has medium-to-high certainty in her evaluation. Bob thinks there is high to very high probability that alternative meets the criterion, but is more uncertain than the others in this assessment Although the prime reason for using Belief Maps is to capture evaluation satisfaction and certainty, they also serve to assist in deliberation. Consider the following:
If information is quantitative, most analyses only generate a single value. However, the uncertainty also needs to be captured. The easiest way to do this is on a simple number line that captures the high low and most likely values. These are incorporated into Accord Professional and Enterprise. 
|
Learn MoreDecision-Maker Blog
|
A Belief Map is a simple graphical representation of the Level of Criterion Satisfaction and Level of Certainty for a particular evaluation. It is, in effect a method for mapping data uncertainty. 