question

Sebastien avatar image
0 Likes"
Sebastien asked Ben Wilson commented

Experimenter replications

Hi,

In our model we used random numbers and distributions. When using the experimenter we obtain a wide range of results between replications, from very bad to very good.

Does the experimenter only change the random number and distributions during a replication or does it do something more ?

We could not find any relevant information on the subject in the manual.

Best regards,

FlexSim 20.1.2
experimenterflexsim 20.1.2replications
· 1
5 |100000

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.

Ben Wilson avatar image Ben Wilson ♦♦ commented ·

Hi @Sebastian, was jason.lightfoot's or Ralf Gruber's answer helpful? If so, please click the red "Accept" button on one of their answers. Or if you still have questions, add a comment and we'll continue the conversation.

If we haven't heard back from you within 3 business days we'll auto-accept an answer, but you can always unaccept and comment back to reopen your question.

0 Likes 0 ·
Jason Lightfoot avatar image
0 Likes"
Jason Lightfoot answered Jason Lightfoot edited

I'm sure you've looked at this tutorial which explains scenarios and replications of an experiment. Basically replications should only change random number stream start seed while scenarios can change variables defined in the experimenter. (Note the format of this will change in 21.0 to model parameters - a Toolbox table)

If you think a replication has a strange result then you can run that interactively by going to the experimenter advanced tab and setting the value of "Repeat Streams of Replication" and see what's going on.


You can change the scenario in the experimenter's scenario tab using the dropdown for "Choose default reset scenario:"


Or in 21.0 beta by selecting the scenario column and pressing "Set model to selected scenario":



1604671642146.png (4.3 KiB)
1604671699796.png (15.0 KiB)
1604671774165.png (11.3 KiB)
1604671874732.png (15.0 KiB)
1604671886000.png (11.3 KiB)
1604671903164.png (4.3 KiB)
5 |100000

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.

Ralf Gruber avatar image
0 Likes"
Ralf Gruber answered

Hello Sebastian,

the Experimenter does only change the random numbers, not the distributions itself. They remain the same with the same parameters for every replication.

Can you post the model, if you do not want it public do it in a private post, please. Otherwise it is tough to guess, what causes the wide spread of replication results...

Best wishes

5 |100000

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.