In this pull request is a Python implementation of the environment for the first case study (Sec. 4.1) in the JellyBean World paper, as per #10.
This new environment is registered under gym as JBW-v1 and JBW-render-v1 for the non-rendering and rendering version of the environment respectively. So to use this environment, simply call
import gym
import jbw
env = gym.make('JBW-v1')
NOTE: there are some slight discrepancies between the configuration mentioned in the paper vs. what was done in the swift implementation. Referencing #10, this implementation uses values from the swift implementation as opposed to what's listed in the paper.
On top of this, I've also included a small change in the Makefile for the visualization - I've referenced the linked libraries after the files to compile due to a referencing issue.
In this pull request is a Python implementation of the environment for the first case study (Sec. 4.1) in the JellyBean World paper, as per #10.
This new environment is registered under
gymasJBW-v1andJBW-render-v1for the non-rendering and rendering version of the environment respectively. So to use this environment, simply callNOTE: there are some slight discrepancies between the configuration mentioned in the paper vs. what was done in the swift implementation. Referencing #10, this implementation uses values from the swift implementation as opposed to what's listed in the paper.
On top of this, I've also included a small change in the Makefile for the visualization - I've referenced the linked libraries after the files to compile due to a referencing issue.