MushroomRL: Simplifying Reinforcement Learning Research

  • 2020-01-09 15:11:21
  • Carlo D'Eramo, Davide Tateo, Andrea Bonarini, Marcello Restelli, Jan Peters
  • 0


MushroomRL is an open-source Python library developed to simplify the processof implementing and running Reinforcement Learning (RL) experiments. Comparedto other available libraries, MushroomRL has been created with the purpose ofproviding a comprehensive and flexible framework to minimize the effort inimplementing and testing novel RL methodologies. Indeed, the architecture ofMushroomRL is built in such a way that every component of an RL problem isalready provided, and most of the time users can only focus on theimplementation of their own algorithms and experiments. The result is a libraryfrom which RL researchers can significantly benefit in the critical phase ofthe empirical analysis of their works. MushroomRL stable code, tutorials anddocumentation can be found at


Quick Read (beta)

loading the full paper ...