Sensitivity Graphs

Assume, for example, that you have developed the following model to calculate the monthly payment for an installment loan.

This model requires three input values: Principal, Annual Rate, and Years. It calculates one output value, Payment.

Next, assume you want to generate a graph showing what happens to Payment as each of the inputs is changed. To do this, point to Sensitivity Analysis in the Tools menu and click Sensitivity Graph. This command brings up the Sensitivity Graph dialog box.

Here, you specify the output and input nodes in your model and the range over which you want to change the input node values. You are not required to explicitly list all input nodes in this dialog box. Instead, you tell DecisionPro to assume that all end nodes are inputs, or that all input nodes have been explicitly declared.

After entering the appropriate options in the Sensitivity Graph dialog box, click OK to produce a graph similar to the following:

This graph shows what happens to the value of the output node, Payment, as each of the input nodes is varied from their current value by -50% to +50%. These results tell you that Payment is most sensitive to changes in Principal. You know this because the Principal line has the steepest slope. You can also see that Payment increases as both Principal and Annual Rate increase, and it decreases as Years increases.

A model can contain tens or even hundreds of end nodes. If this is the case, having DecisionPro treat all end nodes as inputs will create a very cluttered graph. You can explicitly specify which input nodes you want to include in the sensitivity analysis by first declaring the nodes as inputs and then using the each declared input option in the while changing field. Nodes are declared as inputs by choosing the appropriate option in the Type page of the Node Format dialog box.