how to understand Flexsim 2021 optimiser

Question

question

Wang L3 asked May 7, '21 Ben Wilson commented May 19, '21

how to understand Flexsim 2021 optimiser

Dear all,

I have been working on this parking model for a while. With the help and support from this community, I managed to create a dummy model (i.e. a smaller scale model to just test the modeling idea). now I am at the stage to test run the model, and try to interpret the results.

I noticed there is an upgrade in model experimenter in version 2021. In version 2020, we can set variables directly in the experimenter and version 2021 does not allow. As such, I would like to seek your help in advising on how to use the new experimenter.

Back to the problem that I am tring to solve, is that, my management is deciding on whether to provide more seasonal parking to a certain group of staff, and at the same time to study if can close some parking lots due to low utilisation rate.

My idea to implement this in the simulation modeling, is that, to set number of parking lots as a model parameter (as shown in the screenshot below, testing with 50 parking lots vs 52 vs 54, for example). I am also thinking to test the model with different arrival patterns. For example, my current model is having a scheduled arrival with different hourly rates. I might want to increase the hourly arrival by 10%, 20%, and so on.

My question is then, I am actually intersted in testing with two parameters (e.g. number of parking lots, and arrvial rates) at one time, how will flexsim handle this, and how?

Is it done via scenaio testing (i.e. discrete values, 50, 52 or 54 units of parking lots)? or the flexsim will look for a best "answer" in the range that we define (for example, 51 or 53 units of parking lots)?

attached my dummy model as well

SHT model v1.5.fsm

Software Version:

FlexSim 21.1.0

flexsim 21.1.0 optimizer

how-to-set-parameters-to-run-model-for-optimisatio.png (195.2 KiB)

sht-model-v15.fsm (235.1 KiB)

· 1

Ben Wilson ♦♦ commented · May 19, 2021 at 02:18 AM

Hi @Wang L3, was Jörg Vogel's answer helpful? If so, please click the red "Accept" button at the bottom of their answer. 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 ·

Answer 1 · 2021-05-10T04:55:48Z

Joerg Vogel answered May 10, '21

As far as I understand the new input scheme of the experimenter. He works still like the old experimenter. He runs in a scenario a static set of values several times. The difference is, you must parse the values now indirectly as parameters instead of choosing node values. You setup parameters in the parameters tool and in your model you select parameters to replace values or functions.

An optimizer run is still a condition driven run to test the variation of multiple values in multiple test to find an optimum in correlation of values. The conditions are fulfilled if a result lays inside a limited zone or the number different experiments exceeds the limit.

This means you are depending still on OptQuest to find a correlated optimum if you want to test automatically different scenarios values.

question