3. Build an Influence Diagram

Subtitles Enabled

Sign up for a free trial to continue watching this lesson.

Free trial


Using our decision tree, we'll build an influence diagram for our SupraChem problem that will help us structure our final model.

Lesson Notes

Influence diagrams

- Before we jump into building the model, we need to first structure the model
- Influence diagrams are a graphical way of structuring the model
- They are allow you to experiment with different structures without getting mired in detail

Influence diagram elements

- Decision variable: An input we control
- Parameter: Model input we don't control
- Intermediate variable: Calculated from inputs to help us ultimately find our outcome
- Outcome: The value(s) from our model that will provide insight


In the previous lesson, we came up with an equation to satisfy the investment hurdle of SupraChem's executives. We then use the decision tree to come up with a formula for our project cash flows. In this lesson, we'll use these two equations to create an influence diagram for SupraChem's decision on investing in a reference plant. As you'll recall from an earlier course we have four different variable types on our influence diagram. The decision variable, the parameter, the intermediate variable and the outcome. In this particular model, there is no decision variable and the outcome will either be yes or a no. So let's start building our influence diagram. As in the previous modeling course, I like to start with the outcome on the right-hand side of the page which will be a yes, no, decision. If we go back to our first equation, we can see that the decision will be determined by our project cash flows, the five times return requirement, and the initial investment. In our influence diagram, the return requirement is a parameter, the initial investment is a parameter and the project cash flows are an intermediate variable. From our second equation, we can see our project cash flows will be determined by the initial investment, the probability of success and the commercial plant cash flows. The probability of success will be a parameter and the reference plant cost is a parameter. The commercial plant cash flows can be split into the initial setup cost for the commercial plant and the operating cash flows once the plant is up and running. For the operating cash flows, we have revenues which are determined by the market price, the plant capacity and the operating life of the plant. We have production costs, which are also dependent on the plant capacity and the operating life and of course they're dependent on the unit variable costs. Lastly, we have taxes, which will affect the operating cash flows but in the first iteration of this model I'll leave taxes out to keep it as simple as possible. Now that we're happy with our influence diagram, it's time to layout this model in Excel, which I'll do in the next lesson.