Module 3: Decision Trees

Often, we will not know the consequence of a choice until the choice has been made. In other words, we know what might happen, but we do not know what will happen. For example, a manufacturer will be uncertain whether a new product line will succeed. For many problems of this type, decision trees provide an excellent approach.

We start by looking at problem clarification, objective specification, and choice creation. We than study the estimation of outcome probabilities and objective utilities. The importance of sensitivity analysis will be stressed. This will include a discussion and demonstration of Monte Carlo simulation.

Case Study: Phillips Petroleum used decision trees to help them decide how best to allocate the annual exploration budget for drilling projects. This allowed managers to make insightful comparisons of investment opportunities and level of participation in joint drilling projects.
[Ref: M.R. Walls, T. Morahan, J.S. Dyer (1993), Decision Analysis of Exploration Opportunities in the Onshore US at Phillips Petroleum Company, Interfaces, 25, 39-56]