Welcome to Abmarl’s documentation!

Abmarl is a package for developing agent-based simulations and training them with multiagent reinforcement learning. We provide an intuitive command line interface for training, visualizing, and analyzing agent behavior. We define an Agent Based Simulation Interface and Simulation Managers, which control which agents interact with the simulation at each step. We support integration with several popular simulation interfaces, including gym.Env and MultiAgentEnv.

Abmarl is a layer in the Reinforcement Learning stack that sits on top of RLlib. We leverage RLlib’s framework for training agents and extend it to more easily support custom simulations, algorithms, and policies. We enable researchers to rapidly prototype RL experiments and simulation design and lower the barrier for pre-existing projects to prototype RL as a potential solution.