Created
November 7, 2016 14:44
-
-
Save taotao54321/11c9a9eeb0011147aa721bd1fb138dd0 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python3 | |
# -*- coding: utf-8 -*- | |
# Just a handmade algorithm. NOT AI. | |
import sys | |
import gym | |
### 4x4 MAP ### | |
# SFFF | |
# FHFH | |
# FFFH | |
# HFFG | |
############### | |
### 8x8 MAP ### | |
# SFFFFFFF | |
# FFFFFFFF | |
# FFFHFFFF | |
# FFFFFHFF | |
# FFFHFFFF | |
# FHHFFFHF | |
# FHFFHFHF | |
# FFFHFFFG | |
############### | |
# 0:L, 1:D, 2:R, 3:U | |
ACTIONS_MAP = { | |
"4x4" : ( | |
0, 3, 3, 3, | |
0, -1, 0, -1, | |
3, 1, 0, -1, | |
-1, 2, 1, -1, | |
), | |
"8x8" : ( | |
2, 2, 2, 2, 2, 2, 2, 2, | |
3, 3, 3, 3, 3, 3, 3, 2, | |
-1, -1, -1, -1, -1, -1, -1, 2, | |
-1, -1, -1, -1, -1, -1, -1, 2, | |
-1, -1, -1, -1, -1, -1, -1, 2, | |
-1, -1, -1, -1, -1, -1, -1, 2, | |
-1, -1, -1, -1, -1, -1, -1, 2, | |
-1, -1, -1, -1, -1, -1, -1, -1, | |
), | |
} | |
EPISODE_COUNT = 10000 | |
DEBUG = False | |
#DEBUG = True | |
ENVS = { | |
"4x4" : "FrozenLake-v0", | |
"8x8" : "FrozenLake8x8-v0", | |
} | |
def error(msg): | |
sys.exit(msg) | |
def get_action(actions, ob): | |
act = actions[ob] | |
assert 0 <= act <= 3 | |
return act | |
def usage(): | |
error("Usage: FrozenLake-handmade <4x4|8x8> [recdir]") | |
def main(): | |
if len(sys.argv) < 2: usage() | |
env_name = ENVS[sys.argv[1]] | |
actions = ACTIONS_MAP[sys.argv[1]] | |
rec_dir = sys.argv[2] if len(sys.argv) >= 3 else None | |
print("# <{}>".format(env_name)) | |
env = gym.make(env_name) | |
if rec_dir: env.monitor.start(rec_dir) | |
reward_total = 0.0 | |
for episode in range(EPISODE_COUNT): | |
ob = env.reset() | |
if DEBUG: env.render() | |
while True: | |
ob, reward, done, info = env.step(get_action(actions, ob)) | |
if DEBUG: | |
env.render() | |
print(ob, reward, done, info) | |
if done: | |
reward_total += reward | |
break | |
if rec_dir: env.monitor.close() | |
print("episodes: {}".format(EPISODE_COUNT)) | |
print("total reward: {}".format(reward_total)) | |
print("average reward: {:.2f}".format(reward_total / EPISODE_COUNT)) | |
if __name__ == "__main__": main() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment