topic Re: What causes the random errors in the attached model? in FlexSim Forum

What causes the random errors in the attached model?

tee_hiett — Sat, 18 Jul 2020 16:32:51 GMT

[ FlexSim 20.1.0 ]

I am getting a series of five exception error statements occurring randomly in this model. as shown in the graphic, the error occurred once for each first three scenarios. Sometimes it occurs only in one scenario and sometimes it does not occur at all. The messages are shown in the Exceptions-error graphic. when I examine the length of stay for each of the 3 occurrences shown, the length of stay is zero.

Any help would be appreciated.

Thanks,

Tee

Re: What causes the random errors in the attached model?

tanner_p — Mon, 20 Jul 2020 21:42:49 GMT

@tee.hiett,

I took a look at your model and found a couple issues that may be causing the errors you described. The first I found was that the tokens at the "Get NP- Exam-Consult" activity are acquiring a staff resource and then the NP is being released by the parent token. It seems to be throwing things off so that you get the "Bad Task Executer reference" error in the System Console. Maybe it would be better to release the NP in the sub-flow using the Child Token rather than relying on the Parent Token.

Additionally, there are a few errors that occur in the Global Lists. I can't pinpoint what's causing them, but these are the "onFulfill" errors that occur. I would think that resolving some of these errors might prevent problems in the Experimenter results.

Re: What causes the random errors in the attached model?

tee_hiett — Mon, 27 Jul 2020 15:42:01 GMT

Master Clinic Model _18a.fsm

Good Morning @tanner.p

I have modified the Patient Flow chart and I am still getting the same exception messages. When I run the experimenter for 128 replications, I may get 0, 5, 15 or 18 replications where the Avg LOS by Provider is 0.

When I click on the View Results button in the Simulation Experiment Control and examine the Dashboard Statistics for the experiment, all those replications where the LOS is zero, have no value in the charts. When I examine those same replications for the Statistics Tables tab, there are no data in those tables but all of the other replications do have data. So I am assuming those replications with the exception message are not included in any averages.

I have gone about as far as I can go trying to identify any characteristics which would help solve the problem. So I'm turning to the experts for an help.

This master model has almost all the activities we need in any of the 40 odd clinics in this Kirklin Clinic of UAB. I would like to get this problem solved because as soon as we get FlexSim 2020 on the server at the clinic, I plan to use this model to help the members of the performance improvement team use it as a master model for developing models in the clinics.

Thanks for the help.

Tee

PS: I just noticed that the prior version I sent had an subset number error that I fixed

Re: What causes the random errors in the attached model?

tanner_p — Mon, 27 Jul 2020 17:36:55 GMT

Hi @tee.hiett,

I ran the experimenter in your model a few times and was unable to find any replications with the Avg LOS by Provider as 0. Here's a screenshot of what the data looked like for my trials.

Am I looking at the wrong thing? Or maybe I misunderstood the question. Also, is it possible that the Avg LOS by Provider could be 0? Or is that an absolute impossibility based on how your model is set up?

Re: What causes the random errors in the attached model?

tee_hiett — Mon, 27 Jul 2020 19:35:59 GMT

clinic model 1.jpg clinic model 2.jpg clinic model 3.jpg clinic model 4.jpg clinic model 5.jpg

@tanner.p thanks for checking this model. I reran the experimenter with eight replications and looked at the output at each replication. I have enclosed some JPEG files with images from the fifth experiment. I have run the experiment as many as 128 replications with no errors. I have also run the replication hundred and 28 replications with 20 errors. I don't think I have ever had any errors when I run the model manually for one day.

As you well know, this is really frustrating, especially when you don't understand the program in depth begin with :-).

thanks for your help

Re: What causes the random errors in the attached model?

JordanLJohnson — Tue, 28 Jul 2020 17:08:15 GMT

My guess is that this bug is related to the "Restore original state after each replication" checkbox. With that box unchecked, I was able to run the experiment 5 times in a row, and not see any issues. I would try that. I'll be sure to add this bug to the dev list.

Re: What causes the random errors in the attached model?

tee_hiett — Thu, 30 Jul 2020 21:47:27 GMT

@jordan.johnson Unchecking the "Restore original state after each replication" doesn't seem to solve the problem. Any other suggestions?

Thanks for reporting this to the dev list.

Re: What causes the random errors in the attached model?

JordanLJohnson — Thu, 30 Jul 2020 21:54:33 GMT

Hmm. I was unable to replicate the issue without that box checked. I can keep trying, though. But can you confirm that the issue occurs in version 20.1.3? This question is tagged for 20.1.0, and I just want to be sure that the issue exists for you on the latest bug fix release.

Re: What causes the random errors in the attached model?

tee_hiett — Fri, 31 Jul 2020 18:25:51 GMT

Version 20.1.3 did not resolve the problem.

Re: What causes the random errors in the attached model?

JordanLJohnson — Fri, 31 Jul 2020 22:48:46 GMT

Good to know. I'll keep trying to replicate it, but so far, I haven't been able to.

Re: What causes the random errors in the attached model?

tee_hiett — Sat, 01 Aug 2020 21:11:50 GMT

@jordan.johnson Thank for the help. I would suggest you not spend any more time on this now. I have found an old version I was Working on that did not throw that arrow in 128 Replications nor in 256 replications. If I identify the culprit as I rebuilding older version, I will let you know. In fact, I'll let you know the outcome of whatever.

Tee

Re: What causes the random errors in the attached model?

tee_hiett — Fri, 14 Aug 2020 16:40:06 GMT

New Flow Model_3.fsm Sixth experiment.jpg Seventh experiment.jpg eighth experiment.jpg 16 replications.jpg

system console for 16 replications.jpg

@jordan.johnson Eliminating most of the sub flows(New Flow Model_3.fsm) did not eliminate the random problem. I have been running different experiments to find some occurrence that will help with identifying the problem but so far, no luck.

With this new model, I ran 5 experiments without any errors occurring. I have attached the output for the average Length of Stay For the sixth, seventh and eighth experiments. I have also attached Another output for an experiment with 16 replications and the System Console output for that experiment. It doesn't seem to me that the "Restore original state after each replication" has any affect.

I sure would like to get this problem resolved because we will be doing some modeling very soon. Can you think of anything I can be doing or testing?

Thanks for your help.

Tee

PS: I forgot to mention that all of the above tests were done with FlexSim 20.1.3 which I'm using now

Re: What causes the random errors in the attached model?

JordanLJohnson — Mon, 17 Aug 2020 19:47:57 GMT

Hi Tee,

Somehow, there are activities outside of a process flow. If you run this script, you can find them. You can also uncomment the line in the script to delete them:

Array badActivities = [];
forobjectlayerunder(model()) {
	if (isclasstype(a, "ProcessFlow::ProcessBlock")) {
		badActivities.push(a);
	}
}

for (int i = 1; i <= badActivities.length; i++) {
	// uncomment to delete them
        // leave commented to see the list of bad activities
	// badActivities.destroy();
}

return badActivities;

I'm not sure how this happened, and why it doesn't create an issue sooner. But removing these activities may fix your issue.

Re: What causes the random errors in the attached model?

tee_hiett — Tue, 18 Aug 2020 22:11:01 GMT

Thanks, @jordan.johnson. That script got rid of 4 errors. I ran 7 experiments for 256 replications each and got only 1 replication with an average LOS of zero.. Interestingly enough, each one of those errors had the number 73.

Array[4]: /Acquire WR73, /Finish CheckOut73, /Record Notes73, /Remove Patient73. Does that raise any questions in your mind?

Thanks for your help,

Tee