Search found 2395 matches

by tomfid
Wed Nov 14, 2018 1:41 pm
Forum: Vensim
Topic: Changing the order of calculations
Replies: 10
Views: 46

Re: Changing the order of calculations

I think we could say more if we could see a submodel illustrating the problem.
by tomfid
Tue Nov 13, 2018 10:37 pm
Forum: Vensim
Topic: Changing the order of calculations
Replies: 10
Views: 46

Re: Changing the order of calculations

Making the gate position an auxiliary should work. You can store the previous gate position in a stock if you need to preserve the state.
by tomfid
Tue Nov 13, 2018 2:41 pm
Forum: Vensim
Topic: Managing optimization
Replies: 2
Views: 23

Re: Managing optimization

Good question, and yes, it's possible. The result of an optimization is stored in a runname.out file (this is text, so you can look at it in the editor). You can load a .out file the same way you would a changes file (.cin). So, run your optimization in model A, then load optA.out as a change file i...
by tomfid
Sun Nov 11, 2018 8:53 pm
Forum: Vensim
Topic: Forecasting
Replies: 5
Views: 48

Re: Forecasting

By the way, what references said this?
by tomfid
Sun Nov 11, 2018 8:52 pm
Forum: Vensim
Topic: Forecasting
Replies: 5
Views: 48

Re: Forecasting

I shouldn't have said that there was a kernel of truth, because the original statement does not contain one. "SD is not very useful for predicting" is a wholly wrong statement. Whether you want to predict the effect of a policy change, or forecast the state of the system at some point in the future,...
by tomfid
Fri Nov 09, 2018 2:24 am
Forum: Vensim
Topic: Forecasting
Replies: 5
Views: 48

Re: Forecasting

Totally wrong. There is a kernel of truth, which is that forecasting is often not a very high-leverage activity. The "if we know what will happen we'll know what do do" mindset doesn't work in systems with feedback, where actions influence the outcome. However, whether you want to forecast well, or ...
by tomfid
Tue Nov 06, 2018 8:17 pm
Forum: Vensim
Topic: Use File and built-in distributions for sensitivity
Replies: 1
Views: 28

Re: Use File and built-in distributions for sensitivity

No - the two approaches can't be combined. You could generate normal or uniform samples in the file using Excel functions though.
by tomfid
Sat Nov 03, 2018 11:58 pm
Forum: Vensim
Topic: Aging chain with items taken out, on hold and put back
Replies: 20
Views: 181

Re: Aging chain with items taken out, on hold and put back

It is difficult to chose between different solutions, without a deep knowledge of the problem. Exactly. There are lots of approximations that are good enough for some purposes. There's probably no solution for all purposes. Even building an exact representation as a discrete event or agent based si...
by tomfid
Fri Nov 02, 2018 9:59 pm
Forum: Vensim
Topic: Aging chain with items taken out, on hold and put back
Replies: 20
Views: 181

Re: Aging chain with items taken out, on hold and put back

I would generally assume that the first-order approximation is reasonable until proven otherwise. Here's a comparison of 3 approaches. It appears that the dispersion problem in the aging chain is much more significant than the delay order for the hold process. The cohort approach (inspired by Ventit...
by tomfid
Fri Nov 02, 2018 8:49 pm
Forum: Vensim
Topic: Aging chain with items taken out, on hold and put back
Replies: 20
Views: 181

Re: Aging chain with items taken out, on hold and put back

I think the natural heterogeneity of such things negates most of the benefit of the model detail. If you're talking about wind turbines, they're too numerous to worry about it. If coal plants, each one is unique and the duration of retrofit projects has high variance. I can see how the precise timin...
by tomfid
Fri Nov 02, 2018 8:23 pm
Forum: Vensim
Topic: Aging chain with items taken out, on hold and put back
Replies: 20
Views: 181

Re: Aging chain with items taken out, on hold and put back

I'm still not convinced of the policy relevance of preserving every detail of the aging structure.
by tomfid
Fri Nov 02, 2018 2:43 pm
Forum: Vensim
Topic: Aging chain with items taken out, on hold and put back
Replies: 20
Views: 181

Re: Aging chain with items taken out, on hold and put back

Does the delay order for items on hold, or the integer constraint on items, actually matter in some policy-relevant way? If so, it seems like you really need a discrete event simulation.
by tomfid
Fri Nov 02, 2018 2:41 pm
Forum: Vensim
Topic: Aging chain with items taken out, on hold and put back
Replies: 20
Views: 181

Re: Aging chain with items taken out, on hold and put back

You could map the return flow from age x to age x+d of the main stock, right?
by tomfid
Fri Nov 02, 2018 2:29 am
Forum: Vensim
Topic: Aging chain with items taken out, on hold and put back
Replies: 20
Views: 181

Re: Aging chain with items taken out, on hold and put back

I think the difference is simply a matter of focus - this model is the continuous version, that doesn't worry about individual items. I guess you could use an infinite-order delay here, to hold the items by an exact time, and then return them to an appropriate (later) slat in the delay chain. In tha...
by tomfid
Thu Nov 01, 2018 1:26 pm
Forum: Vensim
Topic: Quality Control Question
Replies: 4
Views: 69

Re: Quality Control Question

This model implements defect reduction:
http://metasd.com/2017/11/dynamics-initiative-success/
However, it doesn't have inventories.

For that, you might take a look at how the rework cycle is implemented in a project model:
https://www.vensim.com/documentation/mgu03.htm
by tomfid
Wed Oct 31, 2018 8:41 pm
Forum: Vensim
Topic: Quality Control Question
Replies: 4
Views: 69

Re: Quality Control Question

You can use if-then-else logic, but that's more sensible in a discrete event simulation than in a typical strategic SD model. Specifically, as soon as your stock contains more than one part at a time, you can't really use a true-false choice to decide which ones are defective. More likely, you'll ha...
by tomfid
Tue Oct 30, 2018 5:26 pm
Forum: Vensim
Topic: Payoff report has no info on Goodness of Fit etc
Replies: 4
Views: 86

Re: Payoff report has no info on Goodness of Fit etc

A policy (*P) payoff element has no statistics, because Vensim has no way to know what you're comparing.

If you want the built-in stats, you can use a mixed payoff, with *P for priors and penalty terms, and *C for model-data comparisons.
by tomfid
Tue Oct 30, 2018 4:27 pm
Forum: Vensim
Topic: Payoff report has no info on Goodness of Fit etc
Replies: 4
Views: 86

Re: Payoff report has no info on Goodness of Fit etc

The stats are only present for calibration elements.

I think there was a bug - now fixed - related to synthetic data stats, so check that your version is up to date.

Otherwise, we'd need to see the payoff file at a minimum.
by tomfid
Thu Oct 25, 2018 2:42 pm
Forum: Vensim
Topic: Providing a model like C-ROAD model
Replies: 2
Views: 57

Re: Providing a model like C-ROAD model

You can use Forio or SDeverywhere, too.
https://forio.com/products/epicenter/
http://sdeverywhere.org/
by tomfid
Wed Oct 24, 2018 11:32 am
Forum: Vensim
Topic: Allocate Function
Replies: 4
Views: 66

Re: Allocate Function

by tomfid
Wed Oct 24, 2018 10:18 am
Forum: Vensim
Topic: Allocate Function
Replies: 4
Views: 66

Re: Allocate Function

The help system links some sample models that illustrate this. It works well, as long as the allocated item is a commodity (their are no specific supplier-demander pairing preferences).
by tomfid
Fri Oct 19, 2018 7:34 pm
Forum: Vensim
Topic: unseen variables
Replies: 2
Views: 51

Re: unseen variables

Or you can delete them in the equation editor (there's a delete button at the bottom).
by tomfid
Tue Oct 16, 2018 2:02 pm
Forum: Vensim
Topic: Payoff sensitivity MCMC does not vary with payoff value (1.92, 3.84 or 6.63)
Replies: 2
Views: 77

Re: Payoff sensitivity MCMC does not vary with payoff value (1.92, 3.84 or 6.63)

When you say "results" do you mean the sample points or the reported bounds? I always use the standalone method and evaluate the sample directly these days. The main reason is that the bounds have limited direct utility - if there's any covariance among parameters you need the sample to get the join...
by tomfid
Mon Oct 08, 2018 12:59 pm
Forum: Vensim
Topic: Using different time frames in an on line model
Replies: 3
Views: 90

Re: Using different time frames in an on line model

One possibility would be to use an abstract time axis, where you change the interpretation of time, rather than the time step. However, that seems like it would be (a) messy, and (b) hard to coordinate with the online interface.