Systems simulation comparing different decision rules

Applying the expected utility maximization rule often requires that we quantify probabilities and utilities, an undertaking which can be very complicated, time-consuming, and costly for many decision makers. For consequences of moderate to low utility—where not much is at stake–it's often much more sensible to use a heuristic decision rule This paper considers fourteen different decision rules (eight of them drawn from the election literature) using a simulation. Generally the ranking, approval, weighted ranking, and weighted approval decision rules seem to work well relative to the expected utility rule. In some cases, the rule which ignores all events but the most probable is the best These decision rules were frequently able to get 90% of the utility which would have been acquired using the expected utility rule. Thus, clearly–depending upon the expense of using the expected utility rule versus a simpler heuristic—it may, in fact, be rational for a decision maker to use a heuristic decision rule. This paper provides tables describing how our fourteen decision rules performed under various conditions.