Decision Tree Analysis

last edited by: Moloy Chakraborty on May 11, 2018 4:10 AM login/register to edit this page

 Contents 1 Applications 2 Procedures 3 Instructions 4 Examples 5 Reference A technique that graphically depicts alternate decision paths and assigns probabilities to each path.

Applications

• To facilitate group consensus-based decisions regarding problems, recommended solutions, and/or the evaluation of alternatives during re-engineering.
• To distinguish the alternatives and possible consequences of a decision, or sequence of associated decisions, in advance of final recommendations.

Procedures

1. Isolate decision alternatives.
2. Graphically depicts the decisions in a tree format.
3. Calculate possible financial payoffs, costs, etc., and write them on the diagram.
4. Estimate the probability of the various scenarios and assign payoff variables to calculate expected monetary values for each scenario.
5. See Example.

Instructions

After identifying alternate or candidate solutions, draw a decision tree, depicting the decision alternatives. Use a standard way of depicting the decision and candidate solutions. Lines connect the decisions and the candidate solutions, as depicted in the following example. The lines are labeled to describe the tree.

Enter the costs of each decision and candidate solution on the diagram to calculate the financial returns from each decision. The decisions will have negative numbers, while the candidate solutions should have positive numbers. Sum the numbers to calculate the payoffs. Write the payoffs on the diagram.

Calculate the monetary value by estimating the probability of each candidate solution. The values assigned for probability calculation would sum to 1 if 100% probability can be achieved. Multiply the payoff by the probability to calculate the expected monetary value.

In the example which follows, there are three possible decision paths. For each decision, there are multiple payoffs. The first leg of the tree depicts the cost of choosing that decision, while the second leg of the tree depicts the return for choosing that decision path. The first decision would cost the company \$1,000,000. It has been determined that, if this is the correct action to be taken, there are two possible scenarios for return. The first scenario is a \$1,000,000 return, which leads to a \$0 payoff. The other possibility is a \$5,000,000 return, which leads to a payoff of \$4,000,000. This technique can also be used to supplement cost-benefit analysis (see Cost Estimation, Benefit Estimation, and Cost Benefit Analysis).

Examples

Situation:

PM work for an automobile industry. The organization has initiated a project to upgrade the cars. The sponsor has appointed you as the Project Manager. For this project, one of the requirements is to have special electronic auto parts. Your organization has the capability to produce these auto parts. However, during the planning phase, some stakeholders suggested that the production of the auto parts can be outsourced. But you feel that there is a risk associated with it. Hence you decide to use the decision tree tool to figure out whether it is worth outsourcing or doing it yourself. After brainstorming along with the team, you identify the following values and the risk associated.

Decision ---- Outcome

In-house investment is \$100 Probability 70% Impact Gain is \$120K Probability 30% Impact Loss is \$30K

Outsourcing investment is \$40K Probability 50% Impact Gain is \$100K Probability 50% Impact Loss is \$10K

Question - Use the Decision Tree tool and decide which option is better? Producing in-house or outsourcing

Answer - From the Decision Tree, it is clear that outsourcing is a better option for produce special electronic auto parts.

Reference

1. Alexander Hiam. The Vest-Pocket CEO, Decision-Making Tools for Executives. Prentice Hall, 1990.

last edited by: Moloy Chakraborty on May 11, 2018 4:10 AM login/register to edit this page