Experimenter only works properly for the first batch of child process

Question

question

Hao Zhou asked Aug 2, '16 Adrian Haws edited Aug 9, '16

Experimenter only works properly for the first batch of child process

I have a model that makes experimenter only works properly for the first batch of child processes. There are totally 81 scenarios in the model. For example, my computer has 12 cores. If I set 1 replication per scenario, then during experimentation, scenario 13 replication 1 and all following replications gives me weird results. If I set 2 replications per scenario, then during experimentation, scenario 7 replication 1 and all following replications gives me weird results.

I did checked the model by specified scenario and replication number, it works properly for single run.

Could you help to provide ideas about the reason for this?

Thanks, Hao

Software Version:

FlexSim 16.1.0

experimenter child process restore original state after each replication

· 2

Adrian Haws commented · Aug 03, 2016 at 03:24 PM

@Hao Zhou Would you be able to send your model so we can try to replicate the issue?

0 ·

Hao Zhou Adrian Haws commented · Aug 03, 2016 at 05:55 PM

Hi Adrian,

Thanks for the reply. See attached model and a module that the model relies on.

When you run the experimenter, you can look at a PFM called "GATaktTime(secs)". This value should return a number around 160 since this is the return value of single run. However, depending on how many cores you have in your computer, it returns a number around 130 from the first replication of the second batch of child processes.

Thanks, Hao

model.zip

0 ·

model.zip (2.5 MiB)

Answer 1 · 2016-08-03T21:40:04Z

Sam Stubbs answered Aug 3, '16

I'm trying to replicate the issue, sorry it takes a while on my machine. I'm not entirely sure what you mean by "weird results" could you specify? Also have you tried checking the "Restore original state after each replication" box up on the right side? That might be part of the issue.

· 1

Hao Zhou commented · Aug 03, 2016 at 11:23 PM

Hi @Sam Stubbs,

By saying "weird results" I mean it is a wrong result. As I mentioned in the above comment, it should return a value 160 instead of 130.

I did tried "Restore original state after each replication". It makes experimenter work properly. This checkbox makes the model tree reload for each replication. Does this means I did not reset the model properly? I will double check to see if I do it right or not.

Thank you very much! I really appreciate it.

Hao

0 ·

question