paper: Adding Statement of Need #107
Conversation
| [@gaia] by students and experts alike. | ||
| `mesa-frames` is an extension to the `mesa` [@python-mesa-2020] agent-based modeling framework in Python, designed to enhance performance and scalability for complex simulations involving millions of agents. | ||
|
|
||
| `mesa` has become the most widely used framework for agent-based modeling in Python thanks to its easy-to-use API and object-oriented philosophy. However, iterating through each agent's behavior becomes computationally expensive when the number of agents reaches thousands [@app13010013]. This has led to the development of other frameworks like `Agents.jl` [@Agents.jl], which requires developing in another language, Julia. mesa-frames aims to achieve performance similar to that of `Agents.jl` (NOTE: this needs to be proved) but with a simpler Python syntax. |
There was a problem hiding this comment.
I checked https://arxiv.org/pdf/2101.10072, and there was no evidence that Agents.jl was a reaction to Mesa being slow. Additionally, in the paper:
From our perspective, the biggest take-away of this paper is that Agents.jl is a framework that is simple to use, requiring small amount of written code from the user and overall easy to learn. Despite this, our comparison shows that Agents.jl always exceeds other frameworks in performance, and often also in capability
There was a problem hiding this comment.
You're right there's no consequence of Agents.jl being born out of mesa's speed. I just wanted to remark that the performance of mesa-frames is similar to the one of Agents.jl but the code is in Python.
Maybe it's better to leave out Agents.jl completely?
There was a problem hiding this comment.
Concrete benchmark comparison with Agents.jl is still worth it. They have the official benchmark repo.
There was a problem hiding this comment.
This is very useful. Before implementing the models needed for the benchmark, I would reimplement mesa-frames in Ibis.
This pull request adds the statement of need to the JOSS paper for the
mesa-framesextension.I specified
Ibisinstead ofPolarsandPandassince we are transitioning.Partially addresses #90.