Problem reading/writing tables at end replication

Hi,

i'm having trouble understanding how to reference tables to read/write on end replication trigger on the experimenter.

i'm saving average cycle times of several items on a table, and i want to use another table to calculate at the end of each replication the average of those.

i read the post: http://www.flexsim.com/community/for...ead.php?t=2203 , but i'm not sure i follow... still can't make it work...

regards

Pablo

The quickest way to help you might be with an example model if you have one you can post here.

Originally Posted by Jason Lightfoot

The quickest way to help you might be with an example model if you have one you can post here.

Hi Jason,

i'm ataching a small example.... on the table ctimes i'm counting and calculating the average lenght of stay of each type of item, i want to calculate the average of the counter and the lenght of stay for the number of replications i run on the replication_avg table...

regards

Pablo

The easiest way is to add a PFM for each of the 'average' cells in each row. Then you get a nice report too.

If you don't want that, you use the replication end trigger to pass the table back and update the target table.

Attached are both cases.

Originally Posted by Jason Lightfoot

The easiest way is to add a PFM for each of the 'average' cells in each row. Then you get a nice report too.

If you don't want that, you use the replication end trigger to pass the table back and update the target table.

Attached are both cases.

Thanks! , i knew i could use the pfm on the experimenter, but on the real model i have a lot of different types of products, so i wanted a more direct approach like with the tables...

regards

Pablo

I was getting different results from the PFMs on the experimenter and the tables, so i edited the code from the example you send me so i could get only the average of each type of item on each replication...

when comparing this with the data from the PFMs , there is a lot of difference.... i'm doing something wrong?... why is this different??...

i'm attaching the edited example..

regards

Pablo

I think in my example I added all the samples together to get an overall average which will be different from the PFM.

To match the pfms you just add the average from each replication and divide by the number of replications. I'll post that version in a few mins.

Ah I see now that I've looked at your example.

To compare them you need row 3, column 3 for avg3 PFM (it was row 1 the same as avg 1) and you can also use 'replication' for the column in the end of rep trigger instead of nreps++, to give you a direct comparison for each replication.

Originally Posted by Jason Lightfoot

I think in my example I added all the samples together to get an overall average which will be different from the PFM.

To match the pfms you just add the average from each replication and divide by the number of replications. I'll post that version in a few mins.

Yes, i noticed that, and i changed also (with something like settablenum(table,row,col,(gettablenum(table,row,c ol)*(nreps-1)+gettablenum(tablecopy,row,col))/nreps);.. that i got from another post of yours)... but still i was getting different results, that's why i tried to save all the results from each replication.... still i was getting different results...

regards

Pablo

Originally Posted by Jason Lightfoot

Ah I see now that I've looked at your example.

To compare them you need row 3, column 3 for avg3 PFM (it was row 1 the same as avg 1) and you can also use 'replication' for the column in the end of rep trigger instead of nreps++, to give you a direct comparison for each replication.

still getting different results tho. ?...

they were the same to you?...

regards

Pablo

This model calculates the avg into the table "experiment_avg", while also recording the replication values. You can see they match the PFM statistics.

Yes - there's a problem somewhere - the same replication(s) can generate different results, I think there's a bug.

Seems to have stabilized in this version.

Originally Posted by Jason Lightfoot

Seems to have stabilized in this version.

i'm still getting differences on this version... attaching results of item 1

regards

Pablo

In this one I clear the tables and repaint the view (just to be sure). I've still seen it go wrong though.

Originally Posted by Jason Lightfoot

In this one I clear the tables and repaint the view (just to be sure). I've still seen it go wrong though.

i never see one that it's exactly the same... on this last time on the table replication_avg i get two values that repeat through the experiment per row (rep1=rep3=rep5=rep7=rep9 and rep2=rep4=.. etc)..

regards

Pablo

I can't generate the case with different results today - it's looking correct.

Originally Posted by Jason Lightfoot

I can't generate the case with different results today - it's looking correct.

Had to add the results as PFMs on the experimenter... i never got the same results.

Thanks for the update - I think it deserves another look then. Will post back here if the problem is found.