This is based on CartPole-v0 in OpenAI Gym. I am lazy to give it a better name.
Installation:
$ pip install gym
$ cd /gym-foo/..
$ pip install -e gym-foo
How to get the enviroment:
import gym
env = gym.make('gym_foo:foo-v0')
How to use FooPool to get diverse enviroments:
from gym_foo import FooPool
pool = FooPool(10) # size 10
for env in pool:
# Do something in a new enviroment ...
Default sampling interval of arguments of CartPole simulator:
Argument | Interval | Description |
---|---|---|
gravity | (9.8, 9.8) | gravity constant |
masscart | (0.5, 1.5) | mass of cart |
masspole | (0.05, 0.15) | mass of pole |
mucart | (0.0, 0.5) | friction of cart moving |
mupole | (0.0, 0.5) | friction of pole rotating |
force_mag | (5.0, 15.0) | pushing force |
length | (0.2, 1.2) | length of pole |