Experimenter

[halil_ibrahim_s2]

[ FlexSim 22.1.1 ]

Hello. I'm trying to learn the "experimenter" feature in Flexsim. If I give an example from this model. For example, I want to have a maximum of 20 boxes in the queues. I want to see how many AGVs I need at minimum, using "experimenter" in different scenarios. How do I do this? Or how can I make an experimenter over inter-arrival time. I would be glad if you help.

learning experimenter beginner.fsm

[JordanLJohnson · Accepted solution]

This is a great model for this question. Here's an updated version, with an Experimenter in it:

learning-experimenter-beginner_1.fsm

Here are the steps I used to make this happen:

1. I deleted TaskExecuter2
2. I added TaskExecuter1 to a Group (Group documentation)
3. I added the Queues to a Group
4. I added the Processors to a Group
5. I added a Model parameter called NumAGVs, and I configured it to Delete and Copy objects in the AGVs group. This parameter then controls the number of AGVs in the model. (Model Parameter Documentation)
6. I added two performance measures. One measure how long items on the Processors had to wait to move to the Queue. The other reports the maximum of the maximum content of any of the three queue. (Performance Measure documentation)
7. I opened the Experimenter, and configured the Experiment Job to run 7 scenarios: one Scenario for 1 AGV, 2 AGVs, etc. (Documentation on experiment jobs)

At that point, you can run the Experimenter. For example, here is the plot of Total Block Time for all the scenarios:

It looks like 2 AGVs keep the block time low. More AGVs appear to get in eachother's way, increasing block time.

[halil_ibrahim_s2]

Thank you very much for your return. I understand the model you sent. But I have one more problem. When I uncheck "flush content batween batches" for queues, I can't get block time values (in experimenter). Because I don't want the queue to stay constant all the time. How can I solve this problem?

experimenter.fsm

[JordanLJohnson]

If you uncheck the "Flush content between batches" box, then items can always move to the Queue; they are just part of the next batch. So the Processors won't block, so the total block time will be zero. In that case, your original question makes more sense, in trying to keep the number of boxes below 20. When you run the experiment, you can view the second performance measure:

You can see that with 2 AGVs, none of the queues ever had more than 11 items.