Purpose of the Tornado Diagram add-in
Tornado diagrams graphically display the result of single-factor sensitivity analysis. A typical diagram looks like Figure 1, which shows the effect of four parameters on the result. The uncertainty in the parameter associated with the largest bar, the one at the top of the chart, has the maximum impact on the result, with each successive lower bar having a lesser impact.
 |
The y-axis corresponds to all the uncertain parameters being at their respective nominal, or base, values. For each of the uncertain decision parameters, the chart contains one horizontal bar and two sets of numbers, one of the left and the other to the right of the bar. Each set of numbers corresponds to the result value (upper number) and the value of the parameter at which the result value was reached (the lower number within curly brackets). Negative numbers are shown in parenthesis.
See Common Installation Instructions.
The add-in adds the Tornado Diagram... item under the TM menu.
To create a Tornado diagram , lay out the data so that the nominal (or base) values and the minimum and the maximum values are in adjacent columns as shown in Figure 4 and Figure 5. Ensure that you include the top header row in specifying the data range in the Tornado dialog box.
|
|
 |
 |
The number of samples affects the time required for the analysis as well as the accuracy of the result. Since the software effectively samples the parameter values, increasing the sample size will take longer but will yield more accurate results. Of course, there is no substitute for understanding the problem being studied. For example, if the relationship between the result cell and the parameters in the Data Range is linear, a sample size of 3 will suffice. More complex relationships will require more samples.
The author's email address is web-underscore-contact-at-tushar-hyphen-mehta-dot-cee-oh-em
