Sensitivity Analysis (Vensim)

Hi,

I can see from the chart types and by the appearance of save list and VSC settings that there is the possibility of running sensitivity analysis using a Vensim model in Forio but I can’t find any reference to it in the help documentation. Is there a specific command to run a sensitivity analysis?

Thanks

Lee

Hi,

Thanks for your question (and apologies for the late reply) - here’s instructions on how to add sensitivity analysis for Vensim models to your simulation built with the Interface Builder.

There are two ways of using sensitivity analysis in a interface builder sim.

  • Create a new run, specifically for sensitivity analysis.
  • Run sensitivity analysis on an existing run.

Instructions for creating new sensitivity runs.

  1. Upload .lst and .vsc files from the model settings page
    (See screenshot)

This will internally store (in the model context file) the ‘sensitivityControl’ fields required for sensitivity analysis.

  1. Add a new Sensitivity Analysis (Inputs) page. See screenshot.

  2. Add inputs and click run sensitivity analysis (in preview mode or logged in as a user). This’ll create a new default sensitivity run.

  3. Create a ‘sensitivity outputs’ page. This’ll show the outputs from the most-recently-run sensitivity tun.

With this approach there can only be 1 active sensitivity run per user. You can, of course, rerun sensitivity analysis as many times as you like but the outputs of this run will only show up on the special sensitivity outputs page.

Instructions to run sensitivity analysis on an existing run.

  1. Upload .lst and .vsc files from the model settings page as above.

  2. Create a “regular” run comparison simulation, with standard input and outputs pages.

  3. On any existing ‘Run Analysis’ page you can add/replace a chart with a sensitivity analysis chart (See screenshot).

  4. IF a sensitivity outputs chart exists on that page, each time you ‘analyze’ an existing run by clicking on the radio buttons it’ll auto-run sensitivity analysis based on the inputs for that run.

In technical terms, each run has a 'sensitivityRun: ’ field in it. If you’re on a run analysis page, and have a sensitivity graph on that page the UI will either load the run defined by the sensitivityRunId set on the selected run, or clone the selected run, run analysis on the clone, and sensitivityRun = on the existing run. i.e., it’ll be run at most once per existing run.

Let us know if this helps!

Regards, WILL