tlbtlbtlb
commented
7 years ago From https://gym.openai.com/envs/FrozenLake-v0:
"However, the ice is slippery, so you won't always move in the direction you intend."
Specifically, it has a 1/3 chance of moving in the direction given and 1/3 each to the orthogonal directions. It never moves in the opposite direction. A tabular Q learning agent should have no problem figuring this out.
pablocastilla
Hi!
If I just move in frozen lake to the right twice I don't see the red square in that position.
Quite strange:
Code: import gym import tensorflow as tf import numpy as np
env = gym.make('FrozenLake-v0')
env.reset() rewards = [] for _ in range(2):
state, reward, done, info = env.step(2) env.render()
rewards.append(reward) if done:
env.reset()
Result:
Thanks