RLSim

2 components
SimComponent: Creates a simulation of the Implemented Harvest Game in the Browser
TestComponent: Trains a MultiAgent Reinforcement Learning Environment of the Harvest Game

Run

main.py in the SimComponent. Opens browser window to display the simulation.
GridTestTrain.py in the TestComponent. Will run the Multi Agent Deep Reinforcement Learning Model

Dependencies

mesa, Gym, Ray, tensorboard, numpy

Data

stored in a local directory /ray_results\my_exp.
To view tensorboard graphs for the training enter tensorboard --logdir=~/ray_results\my_exp\experimentfolder into the command prompt.