A Python library for reinforcement learning using Bayesian approaches
MIT License
BayesRL
is a Python library for reinforcement learning using Bayesian
approaches. It stores both agents and environments under separate classes, where
an agent class is a learning algorithm and environments are tasks that the agent
must solve. We include agents and environments for solving and implementing both
Markov decision processes (MDPs) and partially observable Markov decision
processes (POMDPs).
Examples can be found in the directory tests/
. More documentation can be found in the wiki.
To install from pip, run
pip install -e "git+https://github.com/dustinvtran/bayesrl.git#egg=bayesrl"