Ventana software support forum

As far as I recall, it's not possible to scroll the whole screen. An alternative might be to break the screen into several screens, with buttons implementing tab-like switching among them. That might be less overwhelming to the user, too.

This version contains many improvements - it worked really well for us at the SD conference. See "What's New" in the Help system for details. You can download a copy of Ventity from http://vensim.com/ventity.html Feel free to share the link with your colleagues. Download approval is now au...

OK ... here's a solution. There's a glitch: it may not run until you get the next beta (which I'm uploading right now). When I built the original prototype for this, I shouldn't have put the top level in the Model entity, because that's a "singleton," i.e. there can be only one Model in a ...

I'll take a look ...

The next release will improve the export options a lot.

It's in the Help system:
https://www.vensim.com/documentation/sensitivity.htm
Specifically:
https://www.vensim.com/documentation/in ... tivity.htm

The File option does override the number of simulations specified.

Another option is to try symlinking the file in.

I think in some releases there may be a bug in the internal vpm context file management; you could contact us via email for the next build (imminent).

I think we can get you a copy of the Schweppe book for a lot less than $400. It's a good foundation, but not really practical. We've found that the conditions that permit Kalman filtering are rarely met in commercial work. I don't know of a good alternative. Maybe I should write one. The book I've l...

I haven't had a chance to go over the guide in detail, but here are some thoughts. The first variable has approximately a value of 50 and a given weight of 0.25 which gives a value weighted of 12.5 and the second has value of 200 millions and a weight of 30 / 100 millions which gives a value weighte...

The delay duration has units of months. When it's buried in an equation as a number (12), it's effectively dimensionless, so you need to pull it out as a named variable with units, which also has the benefit of exposing the assumption to testing. There were two other implicit parameters in the model...

Sorry it's eating up your time. I'm having a similar issue with Office 365 installation, and it's very frustrating. It's definitely not glitchy for everyone - we have thousands of new users each year, and this particular behavior hasn't come up before. To rename the .ini, start up a command window a...

If you use Vensim to inspect the _outputFileName vdf64, does it look OK? (That would help distinguish a dll data retrieval problem from a model execution/storage problem.)

In sensitivity simulations, you can set the seed to a VECTOR.

See also:

https://www.vensim.com/documentation/noise_seed.htm
https://www.vensim.com/documentation/noise_rng.htm
https://www.vensim.com/documentation/stream_id.htm

You could also flip things around and have one (time->temp) lookup for each latitude, if latitude is in discrete zones. VECTOR LOOKUP might be yet another option if you really need to customize the behavior. Note that you can extract the lookup data from a file with GET XLS or GET DIRECT, if it's al...

I think the behavior differs if your model is an mdl or vmf. Either way, Model>Reform and Clean should force a unit cleanup.

You could use SAMPLE IF TRUE for that too.

If you're using PLE, so you don't have the function, you can store values in a stock. This imposes a delay of one TIME STEP for the integration, but it's not usually a problem.

VECTOR SELECT is another option. In many cases it should be a lot faster than an IF THEN ELSE statement, because it only visits the selected elements. However, I haven't tested to be sure, and it would only matter in a large model. Example: myRange : (r1-r22) active element[myRange] = 0,0,1,0,0,0,1,...

Option 1 gives you a set of named runs that are easier to work with. Option 2 runs very fast.

Some upgrades to this infrastructure are currently underway.

Generically, Option 1 - command script: ! set name and run SIMULATE>RUNNAME|run1 SIMULATE>SETVAL|const1=1.1 MENU>RUN|o ! export to a tab file, using same run name, appending an extra column with :! to add the run name to the description - makes results easy to use in a pivot table MENU>VDF2TAB|!|out...

Also, for just 50 runs, the .dll is probably overkill - too much overhead, assuming you don't need to iterate or make conditional decisions.

I think you could also do it using the Sensitivity File option. Do you have a sample of the input I could take a look at?

There may be one, like switching to a logarithmic space or scaling everything, but it's not obvious to me.

It's a numerical problem. For small sigma, your stock dynamics rely on compounding a very small number - on the order of 1e-44 for sigma=1.42. That's the threshold of failure - for sigma=1.41, the initial stock is ~1e-45, and the compounding fails because the inflow is then too small to add a signif...

One model or two?

Have you tried pasting in both modes (Picture and Replicate)?

I think it's an initialization problem. When mu=1, the density function looks OK, but the cumulative distribution never reaches 1. That's because the stock (Gaussian Density) is initialized to 0. But when x=0 (at time=0), the CDF should be greater than 0, because x=0 is already near the peak of the ...