Describing Decision Problems by Decision Trees
Robert F. Bordley · Wiley Encyclopedia of Operations Research and Management Science · 2010
Abstract Decision trees are especially useful in describing problems where the scientific trial and error processes used in continuous learning are not practical. This article describes the process of formulating the basis for decision trees: values, choices, uncertainties, and deciders. We then describe its application to the simplest problem where we are only concerned with meeting a target and then consider the more general problem—where utility assessment is required—where there is no well‐defined target describing our values. We first present a purely visual approach to decision trees followed by the more precise numerical approach. We close with a discussion of the special case of expected value.