Decision Analysis for Management Judgment, 6th Edition

7Decision trees and influence diagrams

Introduction

When they are first encountered, some decision problems appear overwhelmingly complex. Any attempt at clear thinking can be frustrated by the large number of interrelated elements which are associated with the problem so that, at best, the unaided decision-maker can have only a hazy perception of the issues involved. In these circumstances, decision trees and influence diagrams can be extremely useful in helping people to gain an understanding of the structure of the problems which confront them. Note that the decision trees used here are different from those used in machine learning algorithms for classification. Decision tree classification is applied to problems such as identifying financial transactions as fraudulent or non-fraudulent or predicting which customers might cancel their subscriptions to an internet service, so these types of trees serve a different purpose to those we discuss here.

We have already introduced some very simple decision trees in Chapter 6, but here we will extend the idea to show how multi-stage problems can be modeled. Decision problems are multi-stage in character when the choice of a given option may result in circumstances which will require yet another decision to be made. For example, a company may face an immediate decision relating to the manufacturing capacity which should be provided for a new product. Later, when the product has been on the market for several years, it may have ...