SparkEnv

Simple environments for Reinforcement Learning

1. Robot1d You need import class first "from SparkEnv import Robot1d". Then you need to create environment "env = Robot1d([1, -0.1, -0.3, 0, -0.1, -0.1, 0.5])" where floats are the rewards for robot after moving in this cell. Robot can move left(command "env.step(0)") and right(command "env.step(1)"). The output is positon of robot(list of map's size) and reward of this cell. If the robot tries to move trough the wall the reward is -1. You can render current situation of environment by command "env.render()". After the end of the game you can call the command "env.quit(n)" where n is number of seconds after what the window will close. The sample implementation you can find in the sample.py file.