Software: Vivarium
With members of the Flowers team, we developed Vivarium, a multi-agent simulator in Jax. It serves both research and teaching in Artificial Intelligence and Artificial Life.
This software enables creating simple agents with two motors and two sensors, inspired by Braitenberg Vehicles, in a 2D rigid-body physics world (the physics engine is written in Jax-MD). Despite the simplicity of the agents at an individual level, the interactions between them can lead to complex emergent behaviors and scenarios ! You can easiely launch custom or predefined simulations with config files. The code is still in development, but here is a an example of simulation that can be done at the moment:
The simulations can either run alone, or can be hosted on a server and interacted with in real time from a web interface or Jupyter Notebooks. The web interface mode enables to easiely play with a simulation, and the Notebook case aims to teach students how to program behaviors on these simulated robots with simple Python code ! For example, one can program an eco-evolutionary simulation with preys, predator, ressources and poison, and even evolve the agent’s behaviors with a genetic algorithm.