Sorrel Documentation

Sorrel is a Python API for developing multi-agent reinforcement learning (MARL) environments and agent-based models (ABMs). This framework is intended to function as a flexible platform for computational social scientists to model emergent phenomena among large groups of learning agents.

We include a few basic examples of environments with the package, as well as a set of tutorials (see sidebar) that walk through the necessary steps to build an environment from scratch and wrap a model for use with Sorrel.