Mesa: Agent-based modeling framework in Python3
ฝัง
- เผยแพร่เมื่อ 5 ก.พ. 2025
- This is an overview of Mesa for CoMSES 2018 (www.comses.net....
Resources related to this presentation:
Slides: docs.google.co...
Mesa repo: github.com/pro...
Mesa getting started: mesa.readthedo...
Thanks for creating such wonderful framework, it will have a lot of great use cases.
Do you have any recommendation of how (in terms of overall steps) to implement an optimized (via machine learning or Bayesian) parameter search or validation method? As well, any suggestion for someone who is interested in implementing parallel computing with CPUs/GPUs within your framework? Thanks!
this tool is helping do modeling of the smart electricity networks
Very useful video. Thank you.
It'd be awesome if you made a video tutorialing mesa, like the intro and advanced categories. I've been struggling for a couple of days now...
🥺
it would be useful if you could document Mesa key functions like Netlogo dictionary
Hey! Thank you for your comment! Changes are certainly possible. Can you submit a proposal for discussion via a github issue for a feature request -- github.com/projectmesa/mesa/issues/new/choose?
Hi Jacqueline
I really liked your video about Mesa, is there a site to learn step-by-step how to build ABM models using mesa in jupyter notebooks?
I think this area is being hindered a lot because of the lack of good tools, specially in what concerns parallelization. Technology is progressing super rapidly and most of these projects are outdated. There are some implementations for hpc, but better tools are needed for gpgpu. It has been shown that these models are size sensitive and running them on CPU is definitely not the way to do it.
Look what tensorflow did for machine learning. It is very difficult to partition abm problems efficiently and dwelling into the bowls of gpgpu stack to make it work is not something modelers should be dealing with.
Is there an update on this? I am very interested, please make more videos.
There is no video update, but there is a new paper coming out with SBP-Brims -- sbp-brims.org/2020/
@@JacquelineKazil I don't know what that is, but it sounds like things are working out for you. Glad to hear that.
Too bad that there is no video, because I can't read. (And it spreads the message further ;) )
I am looking forward to the paper, sadly by then my BA deadline is over, and I am stuck with NetLogo for now.
Hi, is there a way I can define different mobility model for the agents? Like the Random Waypoint model for moving the agents? I mean do we have to code the mobility model of our own or is there a predefined mobility model package/function that we can use?
With MESA can I implement the agents or just do simulations?
You can implement the agents. You can see examples here: github.com/projectmesa/mesa/tree/master/examples
@@JacquelineKazil Nice! Watching the video I had the impression that MESA is used only for simulation. I'm glad I'm wrong. I'll look at the examples to identify which one suits me best. Do you know if there is any tutorial for me to follow in the development of my agents?
@@JacquelineKazil Cool, table looks like what I need! I just need to know if he does one more thing, does he support FIPA?
Will I also need to send BDI support, does he have it?