The reinforcement learning library is a library of agents, environment, and experiments that are compatible with RL-Glue and potentially RL-Viz.

All of the components in the RL-Library should be compatible with RL-Glue 3.

Introduction

Check out the new homepage for RL-Library: http://library.rl-community.org

Before you get started, you should really check out the getting started documentation.

We've also added a tutorial about running graphical experiments.

We've started packaging up standalone, compiled versions of the Java projects from RL-Library. Check out the Wiki for links to specific pages.

Current Library Contents

Environments

• MountainCarJava
• TetrisJava
• HelicopterJava
• CartpoleJava
• AcrobotJava

Agents

• RandomAgentJava
• Tile Coding Sarsa Lambda Java

Experiments

• Sample Experiment with RL-Glue in Java
• Sample Graphical Experiment with RLVizApp in Java

There are also other projects available if you checkout the whole project from subversion. They include sample RL-Glue and RL-Viz experiments, a continuous grid world environment, the Octopus environment, a key-board control agent that can be used with RL-Viz, and a few other tidbits. We'll be working on getting them packaged up and getting some wiki documentation ready in the next little while.