I need information regarding best practices in building a large simulation model. In other words, best practices to build a RAM optimized simulation model.
There was this post along the same topic in the FlexSim blog:
https://www.flexsim.com/lessons-from-a-massive-model/
Some other memory-related considerations:
- combine flowitems using join rather than pack whenever possible (minimizes number of flowitems)
- create flowitems only when they are needed and destroy them as soon as possible (minimizes number of flowitems)
- if flowitems do need to persist (or data related to flowitems), consider abstracting flowitems into rows of a bundle and only spawn "physical" 3D flowitems when needed
Other tricks for limiting RAM usage will be more on a case by case basis, depending on what your modeling, the approach you take, the inputs driving the model, and the required outputs.
But in general, a massive number of flowitems is usually where models can become unwieldy. The suggestions above could help.
Is there an way to track the number of flowitems that are currently in play within a model over the course of a simulation run? Is that information stored in some variable? If so, I can write that value to a tracked variable every so often and plot that over time.
Hi @Noah Zenker,
For new questions, please post them as new questions. This thread already has a marked answer.
Please also see the best practices for using this Answers site to better understand the nature of a Question & Answer site, which differs from an old-school forum in regards to having a top question, answers to said question, and related comments.
See points 14 and 15 in particular for tips on the right place to ask a new question.
Thanks!
5 People are following this question.
