I can't stop flip-flopping between #Python and #Netlogo for my #agentBasedModel. Do any #ABM researchers have anything to say about #agentpy and #mesa ?
This is all I can find to compare them...
agentpy.readthedocs.io/en/latest/co...
This is all I can find to compare them...
agentpy.readthedocs.io/en/latest/co...
Comparison — agentpy 0.1.6.dev0 documentation
agentpy.readthedocs.io
May 21, 2025 at 2:49 PM
I can't stop flip-flopping between #Python and #Netlogo for my #agentBasedModel. Do any #ABM researchers have anything to say about #agentpy and #mesa ?
This is all I can find to compare them...
agentpy.readthedocs.io/en/latest/co...
This is all I can find to compare them...
agentpy.readthedocs.io/en/latest/co...
I'm partial to Mesa, but that's mostly because I know & did grad school with most of the developers. I haven't actually used Mesa or AgentPy. Most of my work was done in NetLogo or MASON. However, I do know that Mesa continues to be actively developed & just released v3: mesa.readthedocs.io/stable/
Mesa: Agent-based modeling in Python — Mesa .1 documentation
mesa.readthedocs.io
January 26, 2025 at 1:49 PM
I'm partial to Mesa, but that's mostly because I know & did grad school with most of the developers. I haven't actually used Mesa or AgentPy. Most of my work was done in NetLogo or MASON. However, I do know that Mesa continues to be actively developed & just released v3: mesa.readthedocs.io/stable/
Ah right. Yes, I may eventually switch over to MESA - the interactive visualization is attractive enough and something NetLogo gives you for free! Moving from NetLogo to AgentPy has been an experience in itself. Both handle agent-level updates differently, which makes for interesting outcomes...
January 26, 2025 at 8:42 PM
Ah right. Yes, I may eventually switch over to MESA - the interactive visualization is attractive enough and something NetLogo gives you for free! Moving from NetLogo to AgentPy has been an experience in itself. Both handle agent-level updates differently, which makes for interesting outcomes...
I would be curious about how you're generating your GUI? I haven't got to that stage yet but right now I'm outputting as an animation plot (not that useful in the long term...). Did I miss something obvious in the agentpy documentation?! 😅
January 28, 2025 at 10:38 AM
I would be curious about how you're generating your GUI? I haven't got to that stage yet but right now I'm outputting as an animation plot (not that useful in the long term...). Did I miss something obvious in the agentpy documentation?! 😅
Hi! I'd be curious to learn more about your thoughts on Mesa. I used it a little bit a while ago but have found the programming experience in AgentPy much cleaner & I generally prefer the greater flexibility in GUI (well, it just lets you make your own). I'm also considering Agents.jl to run faster
January 28, 2025 at 10:29 AM
Hi! I'd be curious to learn more about your thoughts on Mesa. I used it a little bit a while ago but have found the programming experience in AgentPy much cleaner & I generally prefer the greater flexibility in GUI (well, it just lets you make your own). I'm also considering Agents.jl to run faster
In NetLogo, an agent runs through each of its functions before the next agent starts.
In AgentPy, all agents iterate through the same function before moving on to the next one.
A trivial difference, but I am learning that this has dire (or interesting??) consequences for my simulation. 😩
In AgentPy, all agents iterate through the same function before moving on to the next one.
A trivial difference, but I am learning that this has dire (or interesting??) consequences for my simulation. 😩
I've been dedicating some time completely rewriting my ABM in Python (from NetLogo), since I hit some limitations with the latter. It's quite cathartic going through the code and sorting out all of the "temporary" fixes thrown in over years of development...😅
January 27, 2025 at 8:17 PM
In NetLogo, an agent runs through each of its functions before the next agent starts.
In AgentPy, all agents iterate through the same function before moving on to the next one.
A trivial difference, but I am learning that this has dire (or interesting??) consequences for my simulation. 😩
In AgentPy, all agents iterate through the same function before moving on to the next one.
A trivial difference, but I am learning that this has dire (or interesting??) consequences for my simulation. 😩
I ended up using AgentPy over MESA. No reason for why I chose one over the other, in fairness.
But AgentPy has its drawbacks, too (or maybe I need it to do more than it offers), so I might end up migrating to MESA after all.
What would you suggest?
agentpy.readthedocs.io/en/latest/
But AgentPy has its drawbacks, too (or maybe I need it to do more than it offers), so I might end up migrating to MESA after all.
What would you suggest?
agentpy.readthedocs.io/en/latest/
AgentPy - Agent-based modeling in Python — agentpy 0.1.6.dev0 documentation
agentpy.readthedocs.io
January 26, 2025 at 7:24 AM
I ended up using AgentPy over MESA. No reason for why I chose one over the other, in fairness.
But AgentPy has its drawbacks, too (or maybe I need it to do more than it offers), so I might end up migrating to MESA after all.
What would you suggest?
agentpy.readthedocs.io/en/latest/
But AgentPy has its drawbacks, too (or maybe I need it to do more than it offers), so I might end up migrating to MESA after all.
What would you suggest?
agentpy.readthedocs.io/en/latest/
May 17, 2025 at 3:00 AM
This results in a modular network structure that’s easier to build, extend, and manage during the simulation. It also works seamlessly with Project Mesa and AgentPy.
3/3
3/3
January 28, 2026 at 3:20 PM
This results in a modular network structure that’s easier to build, extend, and manage during the simulation. It also works seamlessly with Project Mesa and AgentPy.
3/3
3/3