question

Adam_B avatar image
0 Likes"
Adam_B asked Adam_B answered

OptQuest vs. Experimenter

Hello,

I would like to know exactly what I gain by acquiring the optquest add-on. What are the differences between experimenter and optquest?
So far, I have managed to run several simulations where experimenter tested up to 40 combinations, allowing me to determine which one was the most optimal.
I am not yet sure whether experimenter can handle hundreds of combinations or how long it would take—however, based on a proportional estimate from my 40 combinations, processing these scenarios would likely take hours.
This brings me to my question: does optquest operate more efficiently than experimenter in terms of the number of scenarios? and will the processing time be reduced when using optquest?

Regards,

Adam

FlexSim 24.2.3
opt quest
· 1
5 |100000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.

Joerg Vogel avatar image Joerg Vogel commented ·
@Adam_B, OptQuest sets randomly automated values in range of constraints and can finish an independently running setup on its own. The restrictions about the experimenter are applied to OptQuest, too. I mean system memory, speed, stable model setup.
0 Likes 0 ·

2 Answers

Felix Möhlmann avatar image
3 Likes"
Felix Möhlmann answered

The Optimizer is mostly useful when you want/need to find a 'good' set of values for multiple parameters at the same time. The number of possible combinations for just a few parameters can easily reach thousands. So it's often times not feasible to run all of them.

The Optimizer will create a set of initial scenarios (you can also add starting scenarios manually), by choosing parameter values that are equidistant in the allowed range. For example, if you have two parameters with a range between 1 and 100 it might run the following combinations as initial scenarios ([1, 1], [1, 50], [1, 100], [50, 1], [50, 50], [50, 100], [100, 1], [100, 50], [100, 100]). Each scenario is rated by one or multiple objective functions. Those are most commonly maximizing or minimizing a performance measure.

Based on the results of previous iterations, the Optimizer will generate new parameter sets that it thinks are likely to produce good results.

The run speed of each scenario is the same as when using the experimenter. And it likely won't find the single best parameter set. But if there are, for example, a million possible combinations, it can find one that is 'good enough' in a couple hundred iterations.

5 |100000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.

Adam_B avatar image
0 Likes"
Adam_B answered

Thank you for your response. I understand that the Optimizer initially selects scenarios that it believes have the highest potential to achieve the desired outcome. This likely means that it will process significantly fewer scenarios than the Experimenter would in a similar exercise. This approach makes sense in more complex models, where as you mentioned, just a few parameter combinations can turn into thousands of scenarios and system / resources might not be capable to work it out.

5 |100000
Toggle Comment visibility. Current Visibility: Viewable by all users

Up to 12 attachments (including images) can be used with a maximum of 23.8 MiB each and 47.7 MiB total.